WO2021181791A1 - Danger degree calculation device, and device and method for controlling unmanned carrier vehicle - Google Patents

Danger degree calculation device, and device and method for controlling unmanned carrier vehicle Download PDF

Info

Publication number
WO2021181791A1
WO2021181791A1 PCT/JP2020/047022 JP2020047022W WO2021181791A1 WO 2021181791 A1 WO2021181791 A1 WO 2021181791A1 JP 2020047022 W JP2020047022 W JP 2020047022W WO 2021181791 A1 WO2021181791 A1 WO 2021181791A1
Authority
WO
WIPO (PCT)
Prior art keywords
risk
worker
guided vehicle
automatic guided
agv
Prior art date
Application number
PCT/JP2020/047022
Other languages
French (fr)
Japanese (ja)
Inventor
大生 新川
一哲 北角
田中　清明
和哉 浦部
Original Assignee
オムロン株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by オムロン株式会社 filed Critical オムロン株式会社
Publication of WO2021181791A1 publication Critical patent/WO2021181791A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions

Definitions

  • the present invention relates to a technique for calculating the risk of an automatic guided vehicle coming into contact with a worker, and a technique for controlling an automated guided vehicle based on this risk.
  • AGVs Automated guided vehicles that can transport luggage unmanned in factories have a risk of coming into contact with workers while traveling, so the development of AGVs that travel safely is required.
  • Patent Document 1 A technique is known in which a sensor such as LIDAR is provided on the AGV to prevent contact with workers and the like. Further, in Patent Document 1, the distance between the worker and the automatic guided vehicle is calculated by using a wide-angle camera mounted on the ceiling, and if this distance is smaller than the threshold value, the automatic guided vehicle is stopped to avoid contact. Propose a method to do.
  • an object of the present invention is to provide a technique capable of appropriately evaluating the risk of an automatic guided vehicle coming into contact with a worker and controlling the automatic guided vehicle based on an appropriate degree of risk.
  • the present invention adopts the following configuration.
  • the first aspect of the present invention is A detection means for detecting the positions of the worker and the automatic guided vehicle in an environment in which the worker and the automatic guided vehicle move, and With reference to an environmental map including a travelable route of the automatic guided vehicle in the environment, the automatic guided vehicle performs the work based on the distance between the worker and the automatic guided vehicle along the travelable route.
  • a calculation method for calculating the degree of risk given to members, It is a risk calculation device equipped with.
  • An automated guided vehicle is a vehicle that can run unmanned and can transport goods.
  • the guidance system of the automatic guided vehicle is not particularly limited, and an electromagnetic guidance system, an optical guidance system, a magnetic guidance system, an image recognition system, and an autonomous guidance system can be adopted.
  • the automatic guided vehicle can also be controlled by a command from the control device.
  • the worker is a person who exists in an environment (for example, a factory) where an automatic guided vehicle is used. Therefore, in the present invention, "worker” may be read as “person”.
  • the environmental map may include the route on which the automatic guided vehicle can travel in a graph format. That is, the environmental map may have a data format in which the travelable route is represented by a plurality of nodes and links connecting the nodes.
  • the calculation means calculates the risk based on the distance between the worker and the automatic guided vehicle along the travelable route, the worker is more appropriate than calculating the risk based on a simple straight line distance. And the risk of contact with automatic guided vehicles can be evaluated.
  • the "distance between the worker and the automatic guided vehicle along the travelable route” may be obtained as the distance of the shortest route along the transport route between the worker and the automatic guided vehicle. Alternatively, this distance may be obtained as the shortest route among the routes via the node in the traveling direction of the worker and the automatic guided vehicle, whichever is faster.
  • the degree of risk may be calculated to decrease monotonously in a broad sense according to the distance between the worker and the automatic guided vehicle. In other words, the degree of risk may be calculated so that there is a negative correlation with the distance between the worker and the automatic guided vehicle.
  • the calculation means may calculate the degree of risk in consideration of the speed of movement of the worker and the automatic guided vehicle. For example, the calculation means may calculate the degree of risk as the maximum value of the speed of the worker and the speed of the automatic guided vehicle increases. Alternatively, the calculation means may calculate the degree of risk as the relative speed between the worker and the automatic guided vehicle increases.
  • the positions of the worker and the automatic guided vehicle can be detected from the camera image. That is, the risk calculation device of this embodiment further includes a camera that captures an image of the environment, and the detecting means may detect the positions of the worker and the automatic guided vehicle by image processing on the image.
  • the camera is, for example, a fisheye camera (wide-angle camera) mounted on the ceiling, and the number of the cameras may be one or a plurality. If the camera specifications and shooting conditions are known, the position in the real space can be obtained from the position in the image. Highly accurate position detection is possible by using a camera image.
  • the positions of workers and automatic guided vehicles may be detected using a beacon. That is, the risk calculation device of this embodiment further includes a receiver that receives radio waves transmitted from the beacon transmitters of the worker and the automatic guided vehicle, and the detection means works based on the reception result of the receiver.
  • the positions of personnel and automatic guided vehicles may be detected.
  • position detection is possible without being affected by occlusion.
  • the risk calculation device may create an environmental map, or the risk calculation device may acquire and use the environment map created in advance.
  • the risk calculation device creates an environmental map
  • the first is a method of photographing the environment with a camera while traveling the automatic guided vehicle, detecting the automatic guided vehicle from the image, and creating an environmental map based on the detection result.
  • the second method is to run an automatic guided vehicle equipped with a position detecting means and create an environmental map based on the position information transmitted from the automatic guided vehicle.
  • the third is a method of presenting an image of the environment to the user, having the user set a travelable route on the image, and creating the environment map based on this setting.
  • Another aspect of the present invention is an automatic guided vehicle control device including each means of the above-mentioned risk calculation device and a control means for controlling the moving speed of the automatic guided vehicle according to the calculated risk.
  • the control means may, for example, divide the risk level into a plurality of levels and control the automatic guided vehicle to move at a predetermined speed according to the level.
  • the control means may directly transmit the control command to the automatic guided vehicle, or may transmit the control command to the host device that transmits the control command to the automatic guided vehicle.
  • the present invention may be regarded as a risk calculation device having at least a part of the above means, or may be regarded as a risk evaluation device, a contact probability calculation device, or the like. Further, the present invention can be regarded as a method including at least a part of the above processing, a program for realizing such a method, or a recording medium in which the program is recorded non-temporarily. It should be noted that each of the above means and treatments can be combined with each other as much as possible to form the present invention.
  • the risk of the automatic guided vehicle coming into contact with the worker can be evaluated more appropriately than before, and therefore the automatic guided vehicle can be controlled based on an appropriate degree of risk.
  • the figure which shows the application example of the AGV control apparatus which concerns on this invention The figure which shows the structure of the AGV control apparatus which concerns on embodiment.
  • the flowchart of the process performed by the AGV control device The figure explaining an example of the environment map creation process.
  • the figure explaining another example of the environmental cartography process The figure explaining an example of the risk degree calculation process.
  • the figure explaining another example of the risk calculation process The figure explaining an example of the control command generation processing.
  • FIG. 1 is a diagram showing a schematic configuration of an automatic guided vehicle control system 1 to which the present invention is applied to a factory where a worker and an automatic guided vehicle (AGV) work in the same place.
  • AGV automatic guided vehicle
  • the camera 10 is installed on the ceiling of the work place, and the work area including the worker 2 and the AGV 40 is imaged.
  • the image captured by the camera 10 is captured by the control device 20.
  • the control device 20 detects the worker 2 and the AGV 40 from the image, refers to the environment map including the travelable route of the AGV, and obtains the distance between the worker 2 and the AGV 40 along the travelable route.
  • the control device 20 calculates the degree of danger that the AGV 40 gives to the worker 2 based on this distance.
  • the control device 20 obtains the positions 5 and 6 on the travelable route 4 of the worker 2 and the AGV 40 detected from the image, and identifies the route 7 along the travelable route 4 connecting the positions 5 and 6.
  • the route 7 may be determined as the shortest route connecting the positions 5 and 6, or may be determined in consideration of at least one of the moving direction, the moving speed, and the planned moving route of the worker 2 and the AGV40.
  • the control device 20 calculates the degree of risk based on the distance (length) of the route 7. Typically, the shorter the route 7, the higher the risk, and the longer the route 7, the lower the risk.
  • the degree of risk may be calculated in consideration of the moving speed of at least one of the worker 2 and the AGV 40 in addition to the length of the route 7. Typically, the greater the moving speed or the relative speed, the greater the risk is calculated.
  • the control device 20 generates a control command based on the risk level calculated as described above and transmits it to the host device 30.
  • the higher the risk the slower the moving speed of the AGV 40 is controlled.
  • the AGV40 is controlled to move at a predetermined speed according to the level of risk.
  • the host device 30 is a device that transmits a control command to the AGV 40, and generates and transmits an AGV control command based on the command from the control device 20.
  • the risk level is calculated based on the distance along the movable path of the AGV40, not on the linear distance between the worker 2 and the AGV40, so that the risk level can be evaluated more appropriately. .. For example, even if the linear distance between the worker 2 and the AGV 40 is short, the route may be detoured and the risk of contact may actually be low. According to this system, the risk in such a case can be calculated low, so that unnecessary intervention can be prevented when the risk of contact is low.
  • control device 20 transmits the control command to the host device 30, but the control device 20 may transmit only the degree of risk to the host device 30. On the contrary, the control device 20 may directly transmit the control command to the AGV 40 without going through the host device 30.
  • the automatic guided vehicle control system 1 (hereinafter, also referred to as system 1) includes a fisheye camera 10, an AGV control device 20, a host device 30, and an automatic guided vehicle 40 (AGV).
  • system 1 includes a fisheye camera 10, an AGV control device 20, a host device 30, and an automatic guided vehicle 40 (AGV).
  • the fisheye camera 10 is an image pickup device having an optical system including a fisheye lens and an image sensor (an image sensor such as a CCD or CMOS). As shown in FIG. 1, for example, the fisheye camera 10 may be installed on the ceiling of the work place with the optical axis facing vertically downward, and may take an image of the work place in all directions (360 degrees).
  • the fish-eye camera 10 is connected to the control device 20 by wire (USB cable, LAN cable, etc.) or wirelessly (WiFi, etc.), and the image data captured by the fish-eye camera 10 is taken into the control device 20.
  • the image data may be either a monochrome image or a color image, and the resolution, frame rate, and format of the image data are arbitrary. In this embodiment, it is assumed that a color (RGB) image captured at 10 fps (10 images per second) is used.
  • RGB color
  • the fisheye camera 10 used in the system 1 may be only one or a plurality of fisheye cameras 10. When a plurality of fisheye cameras 10 are used, it is preferable that the shooting ranges are partially overlapped so that the images can be stitched together.
  • the automatic guided vehicle 40 is a type of autonomous vehicle, and is a vehicle that can automatically travel without a human being performing a driving operation.
  • the automatic guided vehicle is also referred to as an AGV (Automated Guided Vehicle).
  • the AGV 40 receives, for example, a command of the movement destination and the movement speed from the host device 30, but the AGV 40 autonomously determines the route to the movement destination.
  • the guided vehicle of the automatic guided vehicle is an electromagnetic induction type that moves according to the magnetic field generated by the electric current flowing through the electric wire embedded in the floor, a magnetic induction type that moves according to the magnetic field generated by the magnetic material embedded in the floor, and a laser radar.
  • Optical induction type that moves according to the detection result of the reflector, image recognition method that reads the marker drawn on the floor or ceiling and moves according to the result, autonomous movement using a gyro, acceleration sensor, distance sensor, etc.
  • image recognition method that reads the marker drawn on the floor or ceiling and moves according to the result
  • autonomous movement using a gyro, acceleration sensor, distance sensor, etc.
  • the AGV40 may be configured so that it can be boarded by a human or can be operated by a human as long as it can run unmanned. Further, the mechanism of the AGV 40 is not particularly limited as long as the article can be conveyed, because the article is mounted on the AGV40.
  • the host device 30 is a device that transmits various commands to the AGV 40 by wireless communication.
  • the host device 30 commands the AGV 40 to move to a designated point, or commands the AGV 40 to decelerate or stop.
  • the host device 30 receives a command from the AGV control device 20 and generates and transmits a control command for controlling the AGV 40 according to the command.
  • the host device 30 receives an instruction of the moving speed of the AGV 40 from the AGV control device 20, generates an instruction that can be interpreted by the AGV 40 according to the instruction, and transmits the instruction to the AGV 40.
  • the AGV control device 20 includes one or more processors, a main storage device, an auxiliary storage device, a communication device, an input device, and an output device as hardware components thereof, and the processor executes a computer program to execute the following. Execute various processes. Note that some or all of the processing may be executed by a dedicated hardware circuit.
  • the AGV control device 20 of the present embodiment includes an image acquisition unit 21, an environment map creation unit 22, a map storage unit 23, a detection unit 24, a risk calculation unit 25, and a control command transmission unit 26.
  • the image acquisition unit 21 has a function of capturing image data from the fisheye camera 10.
  • the captured image data is delivered to the environment map creation unit 22 or the detection unit 24 according to the processing phase.
  • the environmental map creation unit 22 creates an environmental map that expresses the travelable route of the AGV 40 in a graph format based on the image data of the fisheye camera 10. Details will be described later, but here, the environment map creation process will be briefly described.
  • a moving image is taken by the fisheye camera 10 while the AGV40 is running on various routes in the environment of the work place.
  • the environmental map creation unit 22 detects and tracks the AGV40 from the moving image created in this way, and acquires the actual movement route of the AGV40.
  • the environmental map creation unit 22 obtains a route that can be determined to be the same route from the actual movement route of the AGV 40, and further expresses the route in a graph format.
  • the map created by the environmental map is stored in the map storage unit 23.
  • the environment map creation unit 22 may create an environment map based on the position information of the AGV40 instead of creating the environment map based on the image of the fisheye camera 10, or the operator (human) for the image may create the environment map.
  • a route (area) on which the AGV 40 can travel may be set based on the range designation.
  • the AGV control device 20 creates an environmental map, but the AGV control device 20 may acquire and use the environment map created in advance from the map storage unit 23 or an external device. I do not care.
  • the detection unit 24 detects the human body and the AGV from the image of the fisheye camera 10. Object detection can be performed by any existing method. For example, the detection unit 24 may detect the human body and the AGV by template matching, or may detect the human body and the AGV using a learning model that has been machine-learned in advance to detect the human body and the AGV. Further, the detection unit 24 may track the human body and the AGV once detected to determine their positions.
  • the risk calculation unit 25 calculates the risk given to the worker 2 by the AGV 40 based on the positions of the worker 2 and the AGV 40 detected by the detection unit 24 and the environment map stored in the map storage unit 23. ..
  • the danger that the AGV40 poses to the worker 2 is typically the contact between the AGV40 and the worker 2, but the AGV40 gets too close to the worker 2 and the AGV40 approaches from the blind spot of the worker 2. There are also dangers other than contact, such as the sudden appearance of AGV40 in front of worker 2.
  • the risk calculation unit 25 obtains the distance between the worker 2 and the route of the AGV40 along the movable route of the AGV40 included in the environmental map, and determines the risk according to this distance.
  • the route between the worker 2 and the AGV 40 may be obtained simply as the shortest route, or may be obtained in consideration of at least one of the moving speed, the moving direction, and the predicted route of the worker 2 and the AGV 40.
  • the risk calculation unit 25 calculates larger as the distance between the obtained worker 2 and the AGV 40 is shorter.
  • the risk calculation unit 25 may further calculate the risk by considering at least one of the moving speed of the worker 2 and the moving speed of the AGV 40.
  • the control command transmission unit 26 controls the moving speed of the AGV 40 according to the calculated risk level. Specifically, the control command transmission unit 26 generates a control command instructing the AGV 40 to move at a predetermined movement speed according to the level of danger, and transmits the control command to the host device 30. As a result, the host device 30 generates and transmits a control command that can be interpreted by the AGV 40, and the AGV 40 moves at a specified moving speed.
  • the control command transmission unit 26 may be configured to directly transmit the control command to the AGV 40.
  • FIG. 3 is an overall flowchart of the processing performed by the AGV control device 20. Hereinafter, the processing performed by the AGV control device 20 will be described in more detail with reference to the drawings.
  • step S11 the AGV control device 20 creates an environmental map.
  • the environment map is map data capable of grasping a route (travelable route) on which the AGV 40 may travel in an environment in which the worker 2 and the AGV 40 work.
  • the travelable route is represented by a graph format, that is, a plurality of nodes and links connecting the nodes.
  • the environment map creation process will be specifically explained with reference to FIG. First, the AGV 40 is driven so as to pass at least once a route on which the AGV 40 may travel in the workplace. At this time, the image is taken by the fisheye camera 10.
  • FIG. 4A is a diagram showing an image 400 taken by the fisheye camera 10.
  • the environment map creation unit 22 obtains the movement locus of the AGV40 by detecting and tracking the AGV40.
  • the rectangle 401 in FIG. 4A is a bounding box showing the detection result, and the center thereof can be regarded as the position of the AGV40.
  • any existing method such as a method using template matching or a method using a learning model (discriminator) generated by machine learning can be adopted.
  • FIG. 4B is a diagram schematically showing a work place 402 photographed by the fisheye camera 10.
  • the thick line 403 is a travelable route of the AGV 40.
  • the rectangle 404 in the figure is in the cell line where the worker 2 works, and is designed so that the AGV 40 does not enter the cell line.
  • FIG. 4C is a diagram illustrating a process of generating graph-format travelable route data from the movement locus of the AGV40 detected from the image.
  • the environment map creation unit 22 generates image data 410 including the trajectory 411 of the AGV 40 in the distorted image by connecting the detection positions of the AGV 40.
  • the environmental map creation unit 22 obtains the image data 420 by expanding the image data 410 in a plane.
  • the locus 411 in the distorted image is developed into locus 421.
  • the plane development may be performed based on the configuration of the optical system of the fisheye camera 10, the configuration of the image sensor, and the imaging conditions.
  • the position in the plane-expanded image and the position in the real space can be converted to each other.
  • the locus of the AGV40 is drawn as one line, but when the AGV40 travels on the same route multiple times, a plurality of loci can be obtained.
  • the environment map creation unit 22 may perform a process of combining routes that are close to each other into one route.
  • the environment map creation unit 22 extracts nodes from the image data 420.
  • a node is a place where a plurality of links are connected, and typically corresponds to a position where the direction of a route changes (turning corner) and a position where a plurality of routes intersect (intersection).
  • the environment map creation unit 22 detects the L shape, the T shape, and the cross shape from the image data 420 by a method such as template matching.
  • 6 nodes 431 are extracted.
  • a link is a trajectory that connects nodes. Note that the node extraction may be performed by the operator (human) manually specifying the node.
  • the environmental map creation unit 22 obtains the link distance (distance between nodes).
  • the link distance is a distance in the real space or a value corresponding to the distance.
  • the position in the plane-developed image and the position in the real space correspond to each other, the distance between two points in the image in the real space can be obtained.
  • the environmental map creation unit 22 creates data including the positions of each node, the connection relationship (link) between the nodes, and the distance of the link, that is, graph-format travelable route data as an environmental map.
  • the created environment map data is stored in the map storage unit 23.
  • the locus of the AGV40 is obtained from the image captured by the fisheye camera 10, but the locus of the AGV40 may be obtained by another method.
  • an environmental map can be created as follows. First, it is the same as the above in that the AGV40 is traveled so as to pass the route on which the AGV40 may travel at least once in the workplace.
  • the AGV 40 is provided with a position detecting means for detecting its own position.
  • An example of a position detecting means is an indoor positioning device. The indoor positioning device performs positioning based on at least one of a BLE beacon, a radio signal such as Wi-Fi and IEMS, an ultrasonic signal, and a visible light signal.
  • the position detecting means is a satellite positioning device (for example, a GPS device) using a satellite signal, and a self-position estimation device using a sensor such as a camera or LIDAR.
  • the environment map creation unit 22 obtains the trajectory of the AGV40 by using the position information transmitted from the AGV40. The processing after the locus of AGV40 is obtained is the same as described above.
  • FIG. 5 shows a travelable route setting screen 500 displayed by the environment map creation unit 22.
  • the setting screen 500 includes an image 501 captured by the fisheye camera 10.
  • an example of displaying a plane-expanded image is shown, but a distorted image may be displayed without plane-expanded.
  • the operator can input the locus 502 on the image 501.
  • the environment map creation unit 22 creates an environment map by performing the processes after the node extraction described above on the locus 502 set by the user.
  • the AGV control device 20 may acquire the environment map created in advance from the map storage unit 23 or an external device and use it.
  • the environment map creation process (step S11) may be performed at a timing before the AGV control device 20 controls the AGV 40.
  • Step S12 and subsequent steps are processes for controlling the AGV40.
  • the processes of steps S12 to S14 are executed every frame or every predetermined frame.
  • step S12 the detection unit 24 detects the worker 2 and the AGV 40 from the image of the fisheye camera 10.
  • Object detection can be performed by any existing method.
  • the detection unit 24 may detect the human body and the AGV by template matching, or may detect the human body and the AGV using a learning model that has been machine-learned in advance to detect the human body and the AGV. Further, the detection unit 24 may track the human body and the AGV once detected to determine their positions.
  • step S13 the risk calculation unit 25 calculates the risk given to the worker 2 by the AGV 40 based on the positions of the worker 2 and the AGV 40. More specifically, the risk calculation unit 25 obtains a route between the worker 2 and the AGV 40 along the travelable route of the AGV 40 on the environmental map, and calculates the risk based on the distance of the route. ..
  • FIG. 6A is a flowchart showing a specific example of the risk calculation process.
  • the distance between worker 2 and AGV40 is determined as a simple shortest path.
  • the risk calculation unit 25 identifies the positions on the graph (link) of the travelable route corresponding to the detection positions of the worker 2 and the AGV 40. If the detection position is on the link, the detection position may be used as it is, but it is assumed that the detection position (particularly the detection position of the worker 2) is not located on the link. In this case, the position on the link closest to the detection position may be used. When the position of the detection position on the link is specified, the position divides the link as a new node. The risk calculation unit 25 obtains the distance of the divided links.
  • the worker 2 determines that the AGV 40 is in a safe area where the AGV 40 does not travel, and does not perform the processing after step S132. , The risk may be determined to be zero.
  • step S132 the risk calculation unit 25 obtains the shortest route on the travelable route between the positions of the worker 2 and the AGV 40.
  • the shortest path is obtained by a known shortest path search algorithm such as Dijkstra's method or Bellman-Ford method.
  • step S133 the risk calculation unit 25 calculates the risk based on the distance of the shortest path obtained in step S132.
  • the degree of risk may be determined according to any definition formula as long as it is obtained as a monotonous decrease function in a broad sense of distance.
  • FIG. 6B shows an example of an image (planned developed) taken by the fisheye camera 10.
  • this image one AGV (60) and two workers (61, 62) are shown, and they are detected by the detection unit 24.
  • FIG. 6C shows a graph of the travelable route included in the environmental map.
  • the graph of the travelable route includes six nodes N1 to N6 and links connecting them.
  • step S131 the position on the link corresponding to the detection position of the AGV and the worker is obtained.
  • Nodes 63 to 65 in FIG. 6C are positions on the link corresponding to the detection positions, respectively.
  • each is also referred to as node A, P1 and P2. Since these nodes divide the existing link, the distance of the divided link can be obtained. For example, the distance between the node A and the node N2 is 2 and the distance between the node A and the node N5 is 1.
  • step S132 the shortest path is obtained for each pair of AGV and worker.
  • the shortest path between the node A (63) and the node P1 (64) is required.
  • node A-node N2-node N1-node P1 is obtained as the shortest path between node A and node P1.
  • the distance of this route is also required.
  • the shortest path between node A (63) and node P2 (65) is determined as node A-node N5-node P2, and its distance is determined.
  • the straight-line distance between the AGV 60 and the workers 61 and 62 is almost the same, but by considering the traveling route, the risk between the AGV 60 and the worker 62 can be determined from the risk between the AGV 60 and the worker 61. Can also be calculated large.
  • FIG. 7A is a flowchart showing another specific example of the risk calculation process.
  • the path between the worker and the AGV is determined in consideration of the moving speed of the worker and the AGV.
  • the moving speed direction and speed
  • the moving speed can be determined from the positions of the worker and the AGV in the most recent multiple frames.
  • FIG. 7B shows the positions of the AGV 60 and the workers 61 and 62 and their moving speeds (arrows 70 to 72).
  • step S131 the risk calculation unit 25 specifies a position on the graph (link) of the travelable route corresponding to the detection position of the worker and the AGV. Since this process is the same as above, the description thereof will be omitted.
  • step S135 the risk calculation unit 25 selects the worker and the AGV, whichever is faster. This selection is made for each worker-AGV pair. For example, in the situation shown in FIG. 7B, the worker 61 is selected for the AGV 60 and the worker 61, and the AGV 60 is selected for the AGV 60 and the worker 62.
  • step S136 the risk calculation unit 25 identifies the node to which the worker or AGV (hereinafter referred to as a high-speed moving object) selected in step S135 will reach next.
  • step S137 the risk calculation unit 25 finds the shortest path between the node identified in step S136 and the worker or AGV (hereinafter referred to as a low-speed moving object) not selected in step S135.
  • step S138 the risk calculation unit 25 determines as a route between the worker and the AGV, which is the sum of the route connecting the position of the high-speed moving object and the next node and the shortest route obtained in step S135. .. That is, in the processing of steps S135 to S138 of this example, among the routes between the worker and the AGV along the travelable route, the shortest route via the node in the traveling direction of the high-speed moving object is determined.
  • the route 75 is obtained as the shortest route between the node N4 and the node A (73) (S137), and the route connecting the node P1 (74) and the node N4 and the route 75 are added together to form the worker 61 and the AGV60. It is determined as a route between (S138).
  • the AGV 60 and the worker 62 will be specifically described with reference to FIG. 7D.
  • the AGV60 since the AGV60 is moving faster, the AGV60 is selected as the high-speed moving object and the worker 62 is selected as the low-speed moving object (S135). Since the AGV60 is moving to the right in the figure, the next node to reach is node N5 (S136). Therefore, the route 78 is obtained as the shortest route between the node N5 and the node P2 (77) (S137), and the route 79 and the route 79 connecting the node A (73) and the node N5 are added to the worker 62 and the AGV60. It is determined as a route between (S138).
  • step S139 the risk calculation unit 25 calculates the risk based on the distance of the route between the worker and the AGV and the moving speed of the worker and the AGV.
  • the degree of risk is inversely proportional to the distance, but the smaller the distance, the greater the degree of risk should be calculated.
  • the greater the relative speed between the worker and the AGV the greater the risk may be calculated.
  • the relative speed is calculated in consideration of the direction of the path connecting the worker and the AGV, and the sum of the speeds is the relative speed when traveling in different directions (opposing directions), and when traveling in the same direction.
  • the difference in speed (absolute value) is the relative speed.
  • the degree of risk may be calculated based only on the distance of the route without considering the speed.
  • step S14 the process of creating and transmitting the AGV control command in step S14 will be described.
  • FIG. 8A is a flowchart showing details of the AGV control command generation / transmission process in step S14.
  • the control command transmission unit 26 compares the risk risk obtained in step S13 with the threshold values TH1 and TH2 (TH1 ⁇ TH2), and determines the traveling speed of the AGV40 based on the comparison result.
  • the control command transmission unit 26 determines the traveling speed to V1 (S142) when the risk level Risk is equal to or less than the first threshold value TH1 (S141-YES), and determines the traveling speed to V1 (S142), and when the risk level Risk is equal to or less than the second threshold value TH2 (S143). -YES) determines the traveling speed to V2 (V2 ⁇ V1) (S144), otherwise (S143-NO) determines the traveling speed to zero (S145). Then, the control command transmission unit 26 generates a control command including the traveling speed determined above and transmits it to the host device 30 (S146).
  • FIG. 8B is a graph showing the relationship between the risk risk and the traveling speed V of the AGV in the control according to the above flowchart.
  • the risk level is divided into three levels according to two threshold values, and the AGV is controlled to move at a predetermined speed (V1, V2, 0) for each level.
  • V1, V2, 0 a predetermined speed
  • the number of levels does not have to be three and may be divided into two levels or four or more levels.
  • the same threshold value is adopted when the speed is increased and when the speed is decreased, but different threshold values may be adopted when the speed is increased and decreased to have hysteresis.
  • control command of the AGV is generated and transmitted for each process, but the control command is generated and transmitted only when it becomes necessary to change the traveling speed of the AGV. It doesn't matter.
  • the risk level is calculated based on the distance of the path along the movable path of the AGV, not on the simple linear distance between the worker and the AGV.
  • the risk of contact between personnel and AGV can be evaluated more appropriately. Therefore, it is possible to avoid a situation in which the AGV is stopped even though the risk of contact is originally low.
  • the AGV control device 20 generates a control command based on the degree of risk and transmits it to the host device 30, and the host device 30 transmits the control command to the AGV 40.
  • the AGV control device 20 may transmit only the degree of risk to the host device 30, and the host device 30 may determine the control of the AGV. Further, the AGV control device 20 may directly transmit a control command to the AGV 40.
  • a risk calculation method performed by the computer (20).

Landscapes

  • Engineering & Computer Science (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

This danger degree calculation device is provided with: a detection means for detecting, in an environment where an operator and an unmanned carrier vehicle move, the positions of the operator and the unmanned carrier vehicle; and a calculation means for calculating, referring to an environment map including a travelable route of the unmanned carrier vehicle in the environment, the degree of danger imposed on the operator by the unmanned carrier vehicle on the basis of the distance between the operator and the unmanned carrier vehicle along the travelable route.

Description

危険度算出装置、無人搬送車の制御装置、および方法Risk calculation device, automatic guided vehicle control device, and method
　本発明は、無人搬送車が作業員と接触する危険度を算出する技術、およびこの危険度に基づいて無人搬送車を制御する技術に関する。 The present invention relates to a technique for calculating the risk of an automatic guided vehicle coming into contact with a worker, and a technique for controlling an automated guided vehicle based on this risk.
　工場内において無人で荷物を搬送可能な無人搬送車（ＡＧＶ）は、走行中に作業員と接触する危険性があるため、安全に走行するＡＧＶの開発が求められている。 Automated guided vehicles (AGVs) that can transport luggage unmanned in factories have a risk of coming into contact with workers while traveling, so the development of AGVs that travel safely is required.
　ＡＧＶにＬＩＤＡＲ等のセンサを設けて、作業員等との接触を防止する技術が知られている。また、特許文献１は、天井に取り付けられた広角カメラを用いて作業員と無人搬送車の間の距離を算出し、この距離が閾値よりも小さければ無人搬送車を停止することで接触を回避する手法を提案する。 A technique is known in which a sensor such as LIDAR is provided on the AGV to prevent contact with workers and the like. Further, in Patent Document 1, the distance between the worker and the automatic guided vehicle is calculated by using a wide-angle camera mounted on the ceiling, and if this distance is smaller than the threshold value, the automatic guided vehicle is stopped to avoid contact. Propose a method to do.
特開平３－１６３６０７号公報Japanese Unexamined Patent Publication No. 3-163607
　しかしながら、特許文献１の手法によると、作業員と無人搬送車の間の距離のみで危険度を判定しているので、実際には接触が生じない場面でも無人搬送車が停止してしまうことがある。また、ＬＩＤＡＲ等のセンサを用いる手法では、センサの検知範囲外の作業員には対処できない。 However, according to the method of Patent Document 1, since the degree of danger is determined only by the distance between the worker and the automatic guided vehicle, the automatic guided vehicle may stop even when no contact actually occurs. be. In addition, a method using a sensor such as LIDAR cannot deal with workers outside the detection range of the sensor.
　そこで、本発明は、無人搬送車が作業員と接触する危険性を従来よりも適切に評価し、適切な危険度に基づいて無人搬送車を制御可能な技術を提供することを目的とする。 Therefore, an object of the present invention is to provide a technique capable of appropriately evaluating the risk of an automatic guided vehicle coming into contact with a worker and controlling the automatic guided vehicle based on an appropriate degree of risk.
　上記目的を達成するために本発明は、以下の構成を採用する。 In order to achieve the above object, the present invention adopts the following configuration.
　本発明の第一側面は、
　作業員および無人搬送車が移動する環境における前記作業員および前記無人搬送車の位置を検出する検出手段と、
　前記環境において前記無人搬送車の走行可能経路を含む環境地図を参照して、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の距離に基づいて、前記無人搬送車が前記作業員に与える危険度を算出する算出手段と、
　を備える、危険度算出装置である。 The first aspect of the present invention is
A detection means for detecting the positions of the worker and the automatic guided vehicle in an environment in which the worker and the automatic guided vehicle move, and
With reference to an environmental map including a travelable route of the automatic guided vehicle in the environment, the automatic guided vehicle performs the work based on the distance between the worker and the automatic guided vehicle along the travelable route. A calculation method for calculating the degree of risk given to members,
It is a risk calculation device equipped with.
　無人搬送車は、無人で走行可能であり、かつ物品を搬送可能な車両である。無人搬送車の誘導方式は特に限定されず、電磁誘導方式、光学誘導方式、磁気誘導方式、画像認識方式、自律誘導方式が採用可能である。無人搬送車は、制御装置からの指令による制御も可能でありうる。 An automated guided vehicle is a vehicle that can run unmanned and can transport goods. The guidance system of the automatic guided vehicle is not particularly limited, and an electromagnetic guidance system, an optical guidance system, a magnetic guidance system, an image recognition system, and an autonomous guidance system can be adopted. The automatic guided vehicle can also be controlled by a command from the control device.
　作業員は、無人搬送車が用いられる環境（例えば、工場）に存在する人間である。したがって、本発明において「作業員」は「人」と読み替えられてもよい。 The worker is a person who exists in an environment (for example, a factory) where an automatic guided vehicle is used. Therefore, in the present invention, "worker" may be read as "person".
　環境地図は、無人搬送車が走行可能な経路をグラフ形式で含んでもよい。すなわち、環境地図は、走行可能経路が複数のノードとノード間を接続するリンクとによって表現されたデータ形式を有してもよい。 The environmental map may include the route on which the automatic guided vehicle can travel in a graph format. That is, the environmental map may have a data format in which the travelable route is represented by a plurality of nodes and links connecting the nodes.
　算出手段は、走行可能経路に沿った作業員と無人搬送車の間の距離に基づいて危険度を算出するので、単純な直線距離に基づいて危険度を算出するよりも、より適切に作業員と無人搬送車の接触の危険度を評価できる。 Since the calculation means calculates the risk based on the distance between the worker and the automatic guided vehicle along the travelable route, the worker is more appropriate than calculating the risk based on a simple straight line distance. And the risk of contact with automatic guided vehicles can be evaluated.
　「走行可能経路に沿った作業員と無人搬送車の間の距離」は、作業員と無人搬送車の間の搬送経路に沿った最短経路の距離として求められてもよい。あるいは、この距離は、作業員と無人搬送車のうち速さが早い方の進行方向にあるノードを経由する経路の中での最短経路として求められてもよい。 The "distance between the worker and the automatic guided vehicle along the travelable route" may be obtained as the distance of the shortest route along the transport route between the worker and the automatic guided vehicle. Alternatively, this distance may be obtained as the shortest route among the routes via the node in the traveling direction of the worker and the automatic guided vehicle, whichever is faster.
　算出手段は、作業員と無人搬送車の間の距離が小さいほど危険度を大きく算出してもよい。言い換えると、危険度は、作業員と無人搬送車の間の距離に応じて広義単調減少するように算出されてもよい。更に別の表現をすると、危険度は、作業員と無人搬送車の間の距離に対して負の相関があるように算出されてもよい。 As the calculation means, the smaller the distance between the worker and the automatic guided vehicle, the greater the degree of risk may be calculated. In other words, the degree of risk may be calculated to decrease monotonously in a broad sense according to the distance between the worker and the automatic guided vehicle. In other words, the degree of risk may be calculated so that there is a negative correlation with the distance between the worker and the automatic guided vehicle.
　算出手段は、作業員と無人搬送車の移動の速さも考慮して危険度を算出してもよい。例えば、算出手段は、作業員の速さと無人搬送車の速さのうちの最大値が大きいほど、危険度を大きく算出してもよい。あるいは、算出手段は、作業員と無人搬送車の相対速度が大きいほど、危険度を大きく算出してもよい。 The calculation means may calculate the degree of risk in consideration of the speed of movement of the worker and the automatic guided vehicle. For example, the calculation means may calculate the degree of risk as the maximum value of the speed of the worker and the speed of the automatic guided vehicle increases. Alternatively, the calculation means may calculate the degree of risk as the relative speed between the worker and the automatic guided vehicle increases.
　本態様において、作業員および無人搬送車の位置はカメラ画像から検出することができる。すなわち、本態様の危険度算出装置は、環境の画像を撮影するカメラをさらに備え、検出手段は、前記画像に対する画像処理によって、作業員および無人搬送車の位置を検出してもよい。カメラは、例えば、天井に取り付けられた魚眼カメラ（広角カメラ）であり、その数は１つであっても複数であってもよい。カメラの仕様と撮影条件が既知であれば、画像中の位置から実空間内の位置を求めることができる。カメラ画像を用いることで高精度な位置検出が可能である。 In this embodiment, the positions of the worker and the automatic guided vehicle can be detected from the camera image. That is, the risk calculation device of this embodiment further includes a camera that captures an image of the environment, and the detecting means may detect the positions of the worker and the automatic guided vehicle by image processing on the image. The camera is, for example, a fisheye camera (wide-angle camera) mounted on the ceiling, and the number of the cameras may be one or a plurality. If the camera specifications and shooting conditions are known, the position in the real space can be obtained from the position in the image. Highly accurate position detection is possible by using a camera image.
　作業員および無人搬送車の位置は、ビーコンを用いて検出してもよい。すなわち、本態様の危険度算出装置は、作業員および無人搬送車が有するビーコン発信器から送信される電波を受信する受信器をさらに備え、検出手段は、受信器の受信結果に基づいて、作業員および無人搬送車の位置を検出してもよい。ビーコンを用いることで、オクルージョンの影響を受けずに位置検出が可能である。 The positions of workers and automatic guided vehicles may be detected using a beacon. That is, the risk calculation device of this embodiment further includes a receiver that receives radio waves transmitted from the beacon transmitters of the worker and the automatic guided vehicle, and the detection means works based on the reception result of the receiver. The positions of personnel and automatic guided vehicles may be detected. By using a beacon, position detection is possible without being affected by occlusion.
　本態様において、環境地図を危険度算出装置が作成してもよいし、あらかじめ作成された環境地図を危険度算出装置が取得して利用してもよい。危険度算出装置が環境地図を作成する場合には、例えば次のような手法がある。第１は、無人搬送車を走行させつつカメラによる環境の撮影を行い、当該画像から無人搬送車を検出し、検出結果に基づいて環境地図を作成する手法である。第２は、位置検出手段を備えた無人搬送車を走行させて、無人搬送車から送信される位置情報に基づいて、環境地図を作成する方法である。第３は、環境を撮影した画像をユーザに提示して、ユーザに走行可能経路を画像上で設定してもらい、この設定に基づいて前記環境地図を作成する手法である。 In this embodiment, the risk calculation device may create an environmental map, or the risk calculation device may acquire and use the environment map created in advance. When the risk calculation device creates an environmental map, for example, there are the following methods. The first is a method of photographing the environment with a camera while traveling the automatic guided vehicle, detecting the automatic guided vehicle from the image, and creating an environmental map based on the detection result. The second method is to run an automatic guided vehicle equipped with a position detecting means and create an environmental map based on the position information transmitted from the automatic guided vehicle. The third is a method of presenting an image of the environment to the user, having the user set a travelable route on the image, and creating the environment map based on this setting.
　本発明の他の態様は、上述の危険度算出装置の各手段と、算出された危険度に応じて無人搬送車の移動速度を制御する制御手段と、を備える無人搬送車の制御装置である。制御手段は、例えば、危険度を複数のレベルに分けて、レベルに応じてあらかじめ定められた速度で無人搬送車が移動するように制御してもよい。制御手段は、無人搬送車に対して制御指令を直接送信してもよいし、無人搬送車に制御指令を送信するホスト装置に対して制御指令を送信してもよい。 Another aspect of the present invention is an automatic guided vehicle control device including each means of the above-mentioned risk calculation device and a control means for controlling the moving speed of the automatic guided vehicle according to the calculated risk. .. The control means may, for example, divide the risk level into a plurality of levels and control the automatic guided vehicle to move at a predetermined speed according to the level. The control means may directly transmit the control command to the automatic guided vehicle, or may transmit the control command to the host device that transmits the control command to the automatic guided vehicle.
　本発明は、上記手段の少なくとも一部を有する危険度算出装置として捉えてもよいし、危険度評価装置、接触確率算出装置などとして捉えてもよい。また、本発明は、上記処理の少なくとも一部を含む方法、または、かかる方法を実現するためのプログラムやそのプログラムを非一時的に記録した記録媒体として捉えることもできる。なお、上記手段および処理の各々は可能な限り互いに組み合わせて本発明を構成することができる。 The present invention may be regarded as a risk calculation device having at least a part of the above means, or may be regarded as a risk evaluation device, a contact probability calculation device, or the like. Further, the present invention can be regarded as a method including at least a part of the above processing, a program for realizing such a method, or a recording medium in which the program is recorded non-temporarily. It should be noted that each of the above means and treatments can be combined with each other as much as possible to form the present invention.
　本発明によれば、無人搬送車が作業員と接触する危険性を従来よりも適切に評価可能であり、したがって、適切な危険度に基づいて無人搬送車を制御できる。 According to the present invention, the risk of the automatic guided vehicle coming into contact with the worker can be evaluated more appropriately than before, and therefore the automatic guided vehicle can be controlled based on an appropriate degree of risk.
本発明に係るＡＧＶ制御装置の適用例を示す図。The figure which shows the application example of the AGV control apparatus which concerns on this invention. 実施形態に係るＡＧＶ制御装置の構成を示す図。The figure which shows the structure of the AGV control apparatus which concerns on embodiment. ＡＧＶ制御装置が行う処理のフローチャート。The flowchart of the process performed by the AGV control device. 環境地図作成処理の一例を説明する図。The figure explaining an example of the environment map creation process. 環境地図作成処理の別の例を説明する図。The figure explaining another example of the environmental cartography process. 危険度算出処理の一例を説明する図。The figure explaining an example of the risk degree calculation process. 危険度算出処理の別の例を説明する図。The figure explaining another example of the risk calculation process. 制御指令生成処理の一例を説明する図。The figure explaining an example of the control command generation processing.
　＜適用例＞
　図１を参照して、本発明の適用例の一つについて説明する。図１は、作業員と無人搬送車（ＡＧＶ）が同じ場所で作業する工場に、本発明を適用した無人搬送車制御システム１の概略構成を示す図である。 <Application example>
One of the application examples of the present invention will be described with reference to FIG. FIG. 1 is a diagram showing a schematic configuration of an automatic guided vehicle control system 1 to which the present invention is applied to a factory where a worker and an automatic guided vehicle (AGV) work in the same place.
　図１の例では、作業場の天井にカメラ１０が設置されており、作業員２やＡＧＶ４０を含む作業領域を撮像する。カメラ１０によって撮像された画像は、制御装置２０に取り込まれる。制御装置２０は、画像から作業員２およびＡＧＶ４０を検出し、ＡＧＶの走行可能経路を含む環境地図を参照して、走行可能経路に沿った作業員２とＡＧＶ４０の間の距離を求める。制御装置２０は、この距離に基づいて、ＡＧＶ４０が作業員２に与える危険度を算出する。 In the example of FIG. 1, the camera 10 is installed on the ceiling of the work place, and the work area including the worker 2 and the AGV 40 is imaged. The image captured by the camera 10 is captured by the control device 20. The control device 20 detects the worker 2 and the AGV 40 from the image, refers to the environment map including the travelable route of the AGV, and obtains the distance between the worker 2 and the AGV 40 along the travelable route. The control device 20 calculates the degree of danger that the AGV 40 gives to the worker 2 based on this distance.
　図１中には、環境地図３と、環境地図３に含まれるＡＧＶの走行可能経路４が模式的に示されている。制御装置２０は、画像から検出された作業員２およびＡＧＶ４０の走行可能経路４上の位置５，６を求め、位置５，６を結ぶ走行可能経路４に沿った経路７を特定する。経路７は、位置５，６を結ぶ最短経路として決定されてもよいし、作業員２およびＡＧＶ４０の移動方向、移動速度、移動予定経路の少なくともいずれかを考慮して決定されてもよい。制御装置２０は、経路７の距離（長さ）に基づいて危険度を算出する。典型的には経路７が短いほど危険度は大きく、経路７が長いほど危険度は小さく算出される。なお、危険度は、経路７の長さ以外に、作業員２およびＡＧＶ４０の少なくとも一方の移動速度も考慮して算出されてもよい。典型的には移動速度あるいは相対速度が大きいほど危険度は大きく算出される。 In FIG. 1, the environmental map 3 and the travelable route 4 of the AGV included in the environmental map 3 are schematically shown. The control device 20 obtains the positions 5 and 6 on the travelable route 4 of the worker 2 and the AGV 40 detected from the image, and identifies the route 7 along the travelable route 4 connecting the positions 5 and 6. The route 7 may be determined as the shortest route connecting the positions 5 and 6, or may be determined in consideration of at least one of the moving direction, the moving speed, and the planned moving route of the worker 2 and the AGV40. The control device 20 calculates the degree of risk based on the distance (length) of the route 7. Typically, the shorter the route 7, the higher the risk, and the longer the route 7, the lower the risk. The degree of risk may be calculated in consideration of the moving speed of at least one of the worker 2 and the AGV 40 in addition to the length of the route 7. Typically, the greater the moving speed or the relative speed, the greater the risk is calculated.
　制御装置２０は、上記のようにして算出した危険度に基づく制御指令を生成してホスト装置３０へ送信する。典型的には、危険度が大きいほど、ＡＧＶ４０の移動速度が遅く制御される。例えば、危険度のレベルに応じてあらかじめ定められた速度で移動するように、ＡＧＶ４０が制御される。 The control device 20 generates a control command based on the risk level calculated as described above and transmits it to the host device 30. Typically, the higher the risk, the slower the moving speed of the AGV 40 is controlled. For example, the AGV40 is controlled to move at a predetermined speed according to the level of risk.
　ホスト装置３０は、ＡＧＶ４０に対して制御指令を送信する装置であり、制御装置２０からの指令に基づいてＡＧＶの制御指令を生成・送信する。 The host device 30 is a device that transmits a control command to the AGV 40, and generates and transmits an AGV control command based on the command from the control device 20.
　このように、本システムにおいては、作業員２とＡＧＶ４０の間の直線距離ではなく、ＡＧＶ４０の移動可能経路に沿った距離に基づいて危険度が算出されるので、危険度をより適切に評価できる。例えば、作業員２とＡＧＶ４０の直線距離が近くても、経路が迂回しており実際には接触の危険性が低い場合もある。本システムによれば、このような場合の危険度を低く算出できるので、接触の危険性が低いときの不要な介入を防止できる。 As described above, in this system, the risk level is calculated based on the distance along the movable path of the AGV40, not on the linear distance between the worker 2 and the AGV40, so that the risk level can be evaluated more appropriately. .. For example, even if the linear distance between the worker 2 and the AGV 40 is short, the route may be detoured and the risk of contact may actually be low. According to this system, the risk in such a case can be calculated low, so that unnecessary intervention can be prevented when the risk of contact is low.
　なお、上記の例では、制御装置２０が制御指令をホスト装置３０に送信しているが、制御装置２０は危険度のみをホスト装置３０に送信してもよい。また逆に、制御装置２０がホスト装置３０を介することなく、制御指令をＡＧＶ４０に対して直接送信しても構わない。 In the above example, the control device 20 transmits the control command to the host device 30, but the control device 20 may transmit only the degree of risk to the host device 30. On the contrary, the control device 20 may directly transmit the control command to the AGV 40 without going through the host device 30.
　＜第１実施形態＞
（構成）
　図２を参照して、本発明の実施形態に係る無人搬送車制御システム１の具体的な構成例を説明する。無人搬送車制御システム１（以下、システム１とも称する）は、魚眼カメラ１０、ＡＧＶ制御装置２０、ホスト装置３０、および無人搬送車４０（ＡＧＶ）を備えている。 <First Embodiment>
(composition)
A specific configuration example of the automatic guided vehicle control system 1 according to the embodiment of the present invention will be described with reference to FIG. The automatic guided vehicle control system 1 (hereinafter, also referred to as system 1) includes a fisheye camera 10, an AGV control device 20, a host device 30, and an automatic guided vehicle 40 (AGV).
［魚眼カメラ］
　魚眼カメラ１０は、魚眼レンズを含む光学系と撮像素子（ＣＣＤやＣＭＯＳなどのイメージセンサ）を有する撮像装置である。魚眼カメラ１０は、例えば図１に示すように、作業場の天井などに、光軸を鉛直下向きにした状態で設置され、作業場の全方位（３６０度）の画像を撮影するとよい。魚眼カメラ１０は制御装置２０に対し有線（ＵＳＢケーブル、ＬＡＮケーブルなど）または無線（ＷｉＦｉなど）で接続され、魚眼カメラ１０で撮影された画像データは制御装置２０に取り込まれる。画像データはモノクロ画像、カラー画像のいずれでもよく、また画像データの解像度やフレームレートやフォーマットは任意である。本実施形態では、１０ｆｐｓ（１秒あたり１０枚）で取り込まれるカラー（ＲＧＢ）画像を用いることを想定している。 [Fisheye camera]
The fisheye camera 10 is an image pickup device having an optical system including a fisheye lens and an image sensor (an image sensor such as a CCD or CMOS). As shown in FIG. 1, for example, the fisheye camera 10 may be installed on the ceiling of the work place with the optical axis facing vertically downward, and may take an image of the work place in all directions (360 degrees). The fish-eye camera 10 is connected to the control device 20 by wire (USB cable, LAN cable, etc.) or wirelessly (WiFi, etc.), and the image data captured by the fish-eye camera 10 is taken into the control device 20. The image data may be either a monochrome image or a color image, and the resolution, frame rate, and format of the image data are arbitrary. In this embodiment, it is assumed that a color (RGB) image captured at 10 fps (10 images per second) is used.
　システム１内で利用される魚眼カメラ１０は、１台のみであってもよいし複数台であってもよい。複数台の魚眼カメラ１０を用いる場合には、撮影範囲を一部重複させて、画像を繋ぎ合わせ可能とするとよい。 The fisheye camera 10 used in the system 1 may be only one or a plurality of fisheye cameras 10. When a plurality of fisheye cameras 10 are used, it is preferable that the shooting ranges are partially overlapped so that the images can be stitched together.
［無人搬送車（ＡＧＶ）］
　無人搬送車４０は、自律走行車両の一種であり、人間が運転操作を行わなくても自動で走行できる搬送車である。以下では、無人搬送車のことをＡＧＶ（Automated Guided Vehicle）とも称する。ＡＧＶ４０は、例えば移動先や移動速度の指令をホスト装置３０から受け付けるが、移動先までの経路はＡＧＶ４０が自律的に決定する。無人搬送車の誘導方式には、床に埋め込まれた電線を流れる電流によって生じる磁場にしたがって移動する電磁誘導式、床に埋め込まれた磁性体によって生じる磁場にしたがって移動する磁気誘導式、レーザーレーダによる反射板の検出結果にしたがって移動する光学誘導式、床や天井に描かれたマーカーを読み取ってその結果にしたがって移動する画像認識方式、ジャイロ・加速度センサ・距離センサなどを用いて自律的に移動する自律誘導方式、およびこれらの組み合わせがあるが、そのいずれの方式を採用してもよい。 [Automated guided vehicle (AGV)]
The automatic guided vehicle 40 is a type of autonomous vehicle, and is a vehicle that can automatically travel without a human being performing a driving operation. Hereinafter, the automatic guided vehicle is also referred to as an AGV (Automated Guided Vehicle). The AGV 40 receives, for example, a command of the movement destination and the movement speed from the host device 30, but the AGV 40 autonomously determines the route to the movement destination. The guided vehicle of the automatic guided vehicle is an electromagnetic induction type that moves according to the magnetic field generated by the electric current flowing through the electric wire embedded in the floor, a magnetic induction type that moves according to the magnetic field generated by the magnetic material embedded in the floor, and a laser radar. Optical induction type that moves according to the detection result of the reflector, image recognition method that reads the marker drawn on the floor or ceiling and moves according to the result, autonomous movement using a gyro, acceleration sensor, distance sensor, etc. There are induction methods and combinations thereof, and any of these methods may be adopted.
　ＡＧＶ４０は、無人で走行可能であれば、人間が搭乗可能に構成されたり、人間によって操作可能に構成されたりしてもよい。また、ＡＧＶ４０は、物品を搬送可能であれば、物品を搭載するため機構は特に限定されない。 The AGV40 may be configured so that it can be boarded by a human or can be operated by a human as long as it can run unmanned. Further, the mechanism of the AGV 40 is not particularly limited as long as the article can be conveyed, because the article is mounted on the AGV40.
［ホスト装置］
　ホスト装置３０は、ＡＧＶ４０に対して種々の指令を無線通信により送信する装置である。例えば、ホスト装置３０は、指定地点に移動するようにＡＧＶ４０に命令したり、減速あるいは停止するようにＡＧＶ４０に命令したりする。本実施形態においては、ホスト装置３０は、ＡＧＶ制御装置２０からの指令を受け取り、その指令にしたがってＡＧＶ４０を制御するための制御指令を生成・送信する。例えば、ホスト装置３０は、ＡＧＶ４０の移動速度の指示をＡＧＶ制御装置２０から受け取り、当該指示に従ってＡＧＶ４０が解釈可能な命令を生成して、ＡＧＶ４０に送信する。 [Host device]
The host device 30 is a device that transmits various commands to the AGV 40 by wireless communication. For example, the host device 30 commands the AGV 40 to move to a designated point, or commands the AGV 40 to decelerate or stop. In the present embodiment, the host device 30 receives a command from the AGV control device 20 and generates and transmits a control command for controlling the AGV 40 according to the command. For example, the host device 30 receives an instruction of the moving speed of the AGV 40 from the AGV control device 20, generates an instruction that can be interpreted by the AGV 40 according to the instruction, and transmits the instruction to the AGV 40.
［ＡＧＶ制御装置］
　ＡＧＶ制御装置２０は、そのハードウェア構成要素として、１つ以上のプロセッサ、主記憶装置、補助記憶装置、通信装置、入力装置、出力装置を備え、プロセッサがコンピュータプログラムを実行することによって、以下の各種処理を実行する。なお、一部又は全部の処理は、専用のハードウェア回路によって実行されてもよい。 [AGV controller]
The AGV control device 20 includes one or more processors, a main storage device, an auxiliary storage device, a communication device, an input device, and an output device as hardware components thereof, and the processor executes a computer program to execute the following. Execute various processes. Note that some or all of the processing may be executed by a dedicated hardware circuit.
　本実施形態のＡＧＶ制御装置２０は、画像取得部２１、環境地図作成部２２、地図記憶部２３、検出部２４、危険度算出部２５、制御指令送信部２６を有している。 The AGV control device 20 of the present embodiment includes an image acquisition unit 21, an environment map creation unit 22, a map storage unit 23, a detection unit 24, a risk calculation unit 25, and a control command transmission unit 26.
　画像取得部２１は、魚眼カメラ１０から画像データを取り込む機能を有する。取り込まれた画像データは、処理フェーズに応じて、環境地図作成部２２または検出部２４に引き渡される。 The image acquisition unit 21 has a function of capturing image data from the fisheye camera 10. The captured image data is delivered to the environment map creation unit 22 or the detection unit 24 according to the processing phase.
　環境地図作成部２２は、魚眼カメラ１０の画像データに基づいて、ＡＧＶ４０の走行可能経路をグラフ形式で表現した環境地図を作成する。詳細は後述するが、ここでは、環境地図作成処理について簡単に説明する。まず、作業場の環境内においてＡＧＶ４０を様々な経路で走行させながら、魚眼カメラ１０による動画像撮影を行う。環境地図作成部２２はこのようにして作成された動画から、ＡＧＶ４０を検出および追跡して、ＡＧＶ４０の実際の移動経路を取得する。環境地図作成部２２は、ＡＧＶ４０の実際の移動経路から同一の経路であると判断できる経路を求めて、さらに経路をグラフ形式で表現する。環境地図によって作成された地図は、地図記憶部２３に記憶される。 The environmental map creation unit 22 creates an environmental map that expresses the travelable route of the AGV 40 in a graph format based on the image data of the fisheye camera 10. Details will be described later, but here, the environment map creation process will be briefly described. First, a moving image is taken by the fisheye camera 10 while the AGV40 is running on various routes in the environment of the work place. The environmental map creation unit 22 detects and tracks the AGV40 from the moving image created in this way, and acquires the actual movement route of the AGV40. The environmental map creation unit 22 obtains a route that can be determined to be the same route from the actual movement route of the AGV 40, and further expresses the route in a graph format. The map created by the environmental map is stored in the map storage unit 23.
　なお、環境地図作成部２２は、魚眼カメラ１０の画像に基づいて環境地図を作成する代わりに、ＡＧＶ４０の位置情報に基づいて環境地図を作成してもよいし、画像に対するオペレータ（人間）の範囲指定に基づいてＡＧＶ４０が走行可能な経路（領域）を設定してもよい。 The environment map creation unit 22 may create an environment map based on the position information of the AGV40 instead of creating the environment map based on the image of the fisheye camera 10, or the operator (human) for the image may create the environment map. A route (area) on which the AGV 40 can travel may be set based on the range designation.
　また、本実施形態では、ＡＧＶ制御装置２０が環境地図を作成しているが、ＡＧＶ制御装置２０は、あらかじめ作成された環境地図を地図記憶部２３あるいは外部の装置から取得して利用しても構わない。 Further, in the present embodiment, the AGV control device 20 creates an environmental map, but the AGV control device 20 may acquire and use the environment map created in advance from the map storage unit 23 or an external device. I do not care.
　検出部２４は、魚眼カメラ１０の画像から、人体およびＡＧＶを検出する。物体検出は既存の任意の手法によって行える。例えば、検出部２４は、テンプレートマッチングによって人体およびＡＧＶを検出してもよいし、人体およびＡＧＶを検出するようにあらかじめ機械学習された学習モデルを用いて人体およびＡＧＶを検出してもよい。また、検出部２４は、一度検出した人体およびＡＧＶを追跡してその位置を求めてもよい。 The detection unit 24 detects the human body and the AGV from the image of the fisheye camera 10. Object detection can be performed by any existing method. For example, the detection unit 24 may detect the human body and the AGV by template matching, or may detect the human body and the AGV using a learning model that has been machine-learned in advance to detect the human body and the AGV. Further, the detection unit 24 may track the human body and the AGV once detected to determine their positions.
　危険度算出部２５は、検出部２４によって検出された作業員２とＡＧＶ４０の位置と、地図記憶部２３に記憶されている環境地図に基づいて、ＡＧＶ４０が作業員２に与える危険度を算出する。ＡＧＶ４０が作業員２に与える危険は、典型的には、ＡＧＶ４０と作業員２との接触であるが、ＡＧＶ４０が作業員２に近づきすぎることや、作業員２の死角からＡＧＶ４０が接近することや、作業員２の前に突然ＡＧＶ４０が現れることなど、接触以外の危険も含まれる。 The risk calculation unit 25 calculates the risk given to the worker 2 by the AGV 40 based on the positions of the worker 2 and the AGV 40 detected by the detection unit 24 and the environment map stored in the map storage unit 23. .. The danger that the AGV40 poses to the worker 2 is typically the contact between the AGV40 and the worker 2, but the AGV40 gets too close to the worker 2 and the AGV40 approaches from the blind spot of the worker 2. There are also dangers other than contact, such as the sudden appearance of AGV40 in front of worker 2.
　危険度算出部２５は、環境地図に含まれるＡＧＶ４０の移動可能経路に沿った、作業員２とＡＧＶ４０の経路の距離を求め、この距離に応じて危険度を決定する。作業員２とＡＧＶ４０の経路は、単純に最短な経路として求められてもよいし、作業員２およびＡＧＶ４０の移動速度、移動方向、予測経路の少なくともいずれかを考慮し求められてもよい。危険度算出部２５は、求められた作業員２とＡＧＶ４０の間の経路の距離が短いほど大きく算出する。危険度算出部２５は、さらに、作業員２の移動の速さおよびＡＧＶ４０の移動の速さの少なくとも一方を考慮して、危険度を算出してもよい。 The risk calculation unit 25 obtains the distance between the worker 2 and the route of the AGV40 along the movable route of the AGV40 included in the environmental map, and determines the risk according to this distance. The route between the worker 2 and the AGV 40 may be obtained simply as the shortest route, or may be obtained in consideration of at least one of the moving speed, the moving direction, and the predicted route of the worker 2 and the AGV 40. The risk calculation unit 25 calculates larger as the distance between the obtained worker 2 and the AGV 40 is shorter. The risk calculation unit 25 may further calculate the risk by considering at least one of the moving speed of the worker 2 and the moving speed of the AGV 40.
　制御指令送信部２６は、算出された危険度に応じて、ＡＧＶ４０の移動速度を制御する。具体的には、制御指令送信部２６は、危険度のレベルに応じてあらかじめ定められた移動速度でＡＧＶ４０が移動するように指示する制御指令を生成し、ホスト装置３０へ送信する。これによりＡＧＶ４０が解釈可能な制御指令をホスト装置３０が生成して送信し、ＡＧＶ４０は指定された移動速度で移動することになる。なお、制御指令送信部２６が、ＡＧＶ４０に対して制御指令を直接送信する構成としても構わない。 The control command transmission unit 26 controls the moving speed of the AGV 40 according to the calculated risk level. Specifically, the control command transmission unit 26 generates a control command instructing the AGV 40 to move at a predetermined movement speed according to the level of danger, and transmits the control command to the host device 30. As a result, the host device 30 generates and transmits a control command that can be interpreted by the AGV 40, and the AGV 40 moves at a specified moving speed. The control command transmission unit 26 may be configured to directly transmit the control command to the AGV 40.
（処理）
　図３は、ＡＧＶ制御装置２０が行う処理の全体フローチャートである。以下、図面を参照して、ＡＧＶ制御装置２０が行う処理をより詳細に説明する。 (process)
FIG. 3 is an overall flowchart of the processing performed by the AGV control device 20. Hereinafter, the processing performed by the AGV control device 20 will be described in more detail with reference to the drawings.
［Ｓ１１：環境地図作成処理］
　ステップＳ１１において、ＡＧＶ制御装置２０は環境地図を作成する。環境地図は、作業員２とＡＧＶ４０が作業を行う環境においてＡＧＶ４０が走行する可能性がある経路（走行可能経路）を把握可能な地図データである。本実施形態において、走行可能経路は、グラフ形式、すなわち、複数のノードとノード間を接続するリンクとによって表現される。 [S11: Environmental map creation process]
In step S11, the AGV control device 20 creates an environmental map. The environment map is map data capable of grasping a route (travelable route) on which the AGV 40 may travel in an environment in which the worker 2 and the AGV 40 work. In the present embodiment, the travelable route is represented by a graph format, that is, a plurality of nodes and links connecting the nodes.
　図４を参照しながら、環境地図作成処理を具体的に説明する。まず、作業場においてＡＧＶ４０が走行する可能性がある経路を少なくとも一度は通るように、ＡＧＶ４０を走行させる。この際、魚眼カメラ１０によって撮像を行う。 The environment map creation process will be specifically explained with reference to FIG. First, the AGV 40 is driven so as to pass at least once a route on which the AGV 40 may travel in the workplace. At this time, the image is taken by the fisheye camera 10.
　図４Ａは魚眼カメラ１０によって撮影される画像４００を示す図である。環境地図作成部２２は、ＡＧＶ４０を検出し追跡することによって、ＡＧＶ４０の移動軌跡を求める。図４Ａ中の矩形４０１は、検出結果を示すバウンディングボックスであり、その中心がＡＧＶ４０の位置と見なせる。ＡＧＶ４０の検出は、テンプレートマッチングを用いた手法や機械学習によって生成した学習モデル（識別器）を用いる手法など既存の任意の手法が採用可能である。 FIG. 4A is a diagram showing an image 400 taken by the fisheye camera 10. The environment map creation unit 22 obtains the movement locus of the AGV40 by detecting and tracking the AGV40. The rectangle 401 in FIG. 4A is a bounding box showing the detection result, and the center thereof can be regarded as the position of the AGV40. For the detection of AGV40, any existing method such as a method using template matching or a method using a learning model (discriminator) generated by machine learning can be adopted.
　図４Ｂは魚眼カメラ１０によって撮影される作業場４０２を模式的に示した図である。太線４０３は、ＡＧＶ４０の走行可能経路である。ここでは、走行可能経路が幅を持たないように示しているが、走行可能経路はある程度の幅を持つ。図中の矩形４０４は、作業員２が作業を行うセルラインの中であり、この中にはＡＧＶ４０は進入しないように設計されている。 FIG. 4B is a diagram schematically showing a work place 402 photographed by the fisheye camera 10. The thick line 403 is a travelable route of the AGV 40. Here, it is shown that the travelable route does not have a width, but the travelable route has a certain width. The rectangle 404 in the figure is in the cell line where the worker 2 works, and is designed so that the AGV 40 does not enter the cell line.
　図４Ｃは、画像から検出されるＡＧＶ４０の移動軌跡からグラフ形式の走行可能経路データを生成する処理を説明する図である。環境地図作成部２２は、ＡＧＶ４０の検出位置を繋ぎ合わせることによって、歪曲した画像内でのＡＧＶ４０の軌跡４１１を含む画像データ４１０を生成する。 FIG. 4C is a diagram illustrating a process of generating graph-format travelable route data from the movement locus of the AGV40 detected from the image. The environment map creation unit 22 generates image data 410 including the trajectory 411 of the AGV 40 in the distorted image by connecting the detection positions of the AGV 40.
　環境地図作成部２２は、画像データ４１０を平面展開することによって、画像データ４２０を得る。歪曲した画像内での軌跡４１１は、軌跡４２１に展開される。平面展開は、魚眼カメラ１０の光学系や画像センサの構成および撮影条件に基づいて行えばよい。平面展開された画像中の位置と現実空間内の位置は、互いに変換可能である。 The environmental map creation unit 22 obtains the image data 420 by expanding the image data 410 in a plane. The locus 411 in the distorted image is developed into locus 421. The plane development may be performed based on the configuration of the optical system of the fisheye camera 10, the configuration of the image sensor, and the imaging conditions. The position in the plane-expanded image and the position in the real space can be converted to each other.
　ここでは、ＡＧＶ４０の軌跡を１本の線として描いているが、同じ経路をＡＧＶ４０が複数回走行した場合には複数本の軌跡が得られる。この場合、環境地図作成部２２は、互いに距離が近い経路を１つの経路としてまとめる処理を施してもよい。 Here, the locus of the AGV40 is drawn as one line, but when the AGV40 travels on the same route multiple times, a plurality of loci can be obtained. In this case, the environment map creation unit 22 may perform a process of combining routes that are close to each other into one route.
　環境地図作成部２２は、画像データ４２０からノードを抽出する。ノードは複数のリンクが接続する場所であり、典型的には、経路の向きが変わる位置（曲がり角）、複数の経路が交差する位置（交差点）が該当する。具体的には、環境地図作成部２２は、画像データ４２０からＬ形状、Ｔ形状、十字形状をテンプレートマッチング等の手法によって検出する。この例では、６つのノード４３１が抽出される。ノード間を接続する軌跡がリンクである。なお、ノードの抽出は、オペレータ（人間）がノードを手動で指定することによって行われてもよい。 The environment map creation unit 22 extracts nodes from the image data 420. A node is a place where a plurality of links are connected, and typically corresponds to a position where the direction of a route changes (turning corner) and a position where a plurality of routes intersect (intersection). Specifically, the environment map creation unit 22 detects the L shape, the T shape, and the cross shape from the image data 420 by a method such as template matching. In this example, 6 nodes 431 are extracted. A link is a trajectory that connects nodes. Note that the node extraction may be performed by the operator (human) manually specifying the node.
　環境地図作成部２２は、リンクの距離（ノード間の距離）を求める。ここで、リンクの距離とは、現実空間内での距離、またはそれに応じた値である。上述したように、平面展開した画像内の位置と現実空間内の位置は対応付いているので、画像中の２点間の現実空間内での距離を求めることができる。 The environmental map creation unit 22 obtains the link distance (distance between nodes). Here, the link distance is a distance in the real space or a value corresponding to the distance. As described above, since the position in the plane-developed image and the position in the real space correspond to each other, the distance between two points in the image in the real space can be obtained.
　環境地図作成部２２は、それぞれのノードの位置と、ノード間の接続関係（リンク）と、リンクの距離とを含むデータ、すなわちグラフ形式の走行可能経路データを環境地図として作成する。作成された環境地図データは、地図記憶部２３に記憶される。 The environmental map creation unit 22 creates data including the positions of each node, the connection relationship (link) between the nodes, and the distance of the link, that is, graph-format travelable route data as an environmental map. The created environment map data is stored in the map storage unit 23.
　環境地図作成処理の変形例を説明する。上記の説明では、魚眼カメラ１０によって撮像した画像からＡＧＶ４０の軌跡を求めているが、その他の手法によってＡＧＶ４０の軌跡を求めてもよい。例えば、次のようにして環境地図を作成できる。まず、作業場においてＡＧＶ４０が走行する可能性がある経路を少なくとも一度は通るように、ＡＧＶ４０を走行させる点は上記と同様である。ここで、ＡＧＶ４０は自らの位置を検出する位置検出手段を備えるものとする。位置検出手段の例は、屋内測位装置である。屋内測位装置は、ＢＬＥビーコンやＷｉ－ＦｉやＩＭＥＳなどの無線信号、超音波信号、可視光信号の少なくともいずれかに基づいて測位を行う。位置検出手段の他の例は、衛星信号を用いた衛星測位装置（例えばＧＰＳ装置）、およびカメラやＬＩＤＡＲなどのセンサを用いた自己位置推定装置である。本変形例では、環境地図作成部２２は、ＡＧＶ４０から送信される位置情報を利用してＡＧＶ４０の軌跡を求める。ＡＧＶ４０の軌跡が得られた後の処理は、上記と同様である。 Explain a modified example of the environmental map creation process. In the above description, the locus of the AGV40 is obtained from the image captured by the fisheye camera 10, but the locus of the AGV40 may be obtained by another method. For example, an environmental map can be created as follows. First, it is the same as the above in that the AGV40 is traveled so as to pass the route on which the AGV40 may travel at least once in the workplace. Here, the AGV 40 is provided with a position detecting means for detecting its own position. An example of a position detecting means is an indoor positioning device. The indoor positioning device performs positioning based on at least one of a BLE beacon, a radio signal such as Wi-Fi and IEMS, an ultrasonic signal, and a visible light signal. Another example of the position detecting means is a satellite positioning device (for example, a GPS device) using a satellite signal, and a self-position estimation device using a sensor such as a camera or LIDAR. In this modification, the environment map creation unit 22 obtains the trajectory of the AGV40 by using the position information transmitted from the AGV40. The processing after the locus of AGV40 is obtained is the same as described above.
　環境地図作成処理の別の変形例は、作業環境を撮影した画像上で、オペレータにＡＧＶ４０の走行可能経路を入力（設定）させる方式である。図５は、環境地図作成部２２が表示する、走行可能経路の設定画面５００を示す。設定画面５００には、魚眼カメラ１０によって撮影された画像５０１が含まれる。ここでは平面展開された画像を表示する例を示しているが、平面展開せずに歪曲画像を表示してもよい。この設定画面５００では、オペレータが画像５０１上に軌跡５０２を入力可能である。決定ボタン５０３が押されると、環境地図作成部２２はユーザによって設定された軌跡５０２に対して、上記で説明したノード抽出以降の処理を施して環境地図を作成する。 Another modification of the environment map creation process is a method in which the operator inputs (sets) the travelable route of the AGV40 on the captured image of the work environment. FIG. 5 shows a travelable route setting screen 500 displayed by the environment map creation unit 22. The setting screen 500 includes an image 501 captured by the fisheye camera 10. Here, an example of displaying a plane-expanded image is shown, but a distorted image may be displayed without plane-expanded. On the setting screen 500, the operator can input the locus 502 on the image 501. When the enter button 503 is pressed, the environment map creation unit 22 creates an environment map by performing the processes after the node extraction described above on the locus 502 set by the user.
　また、ここではＡＧＶ制御装置２０が環境地図を作成しているが、ＡＧＶ制御装置２０は、あらかじめ作成された環境地図を地図記憶部２３あるいは外部の装置から取得して利用しても構わない。 Further, although the AGV control device 20 creates the environment map here, the AGV control device 20 may acquire the environment map created in advance from the map storage unit 23 or an external device and use it.
　環境地図作成処理（ステップＳ１１）は、ＡＧＶ制御装置２０がＡＧＶ４０の制御を行うよりも前のタイミングで行われてよい。ステップＳ１２以降が、ＡＧＶ４０の制御を行う処理である。ステップＳ１２～Ｓ１４の処理は、１フレームごとに、あるいは所定フレームおきに実行される。 The environment map creation process (step S11) may be performed at a timing before the AGV control device 20 controls the AGV 40. Step S12 and subsequent steps are processes for controlling the AGV40. The processes of steps S12 to S14 are executed every frame or every predetermined frame.
［Ｓ１２：作業員およびＡＧＶの検出処理］
　ステップＳ１２において、検出部２４は、魚眼カメラ１０の画像から作業員２とＡＧＶ４０を検出する。物体検出は既存の任意の手法によって行える。例えば、検出部２４は、テンプレートマッチングによって人体およびＡＧＶを検出してもよいし、人体およびＡＧＶを検出するようにあらかじめ機械学習された学習モデルを用いて人体およびＡＧＶを検出してもよい。また、検出部２４は、一度検出した人体およびＡＧＶを追跡してその位置を求めてもよい。 [S12: Worker and AGV detection process]
In step S12, the detection unit 24 detects the worker 2 and the AGV 40 from the image of the fisheye camera 10. Object detection can be performed by any existing method. For example, the detection unit 24 may detect the human body and the AGV by template matching, or may detect the human body and the AGV using a learning model that has been machine-learned in advance to detect the human body and the AGV. Further, the detection unit 24 may track the human body and the AGV once detected to determine their positions.
［Ｓ１３：危険度算出処理］
　ステップＳ１３において、危険度算出部２５は、作業員２とＡＧＶ４０の位置と、に基づいて、ＡＧＶ４０が作業員２に与える危険度を算出する。より具体的には、危険度算出部２５は、環境地図におけるＡＧＶ４０の走行可能経路に沿った、作業員２とＡＧＶ４０の間の経路を求めて、当該経路の距離に基づいて危険度を算出する。 [S13: Risk calculation process]
In step S13, the risk calculation unit 25 calculates the risk given to the worker 2 by the AGV 40 based on the positions of the worker 2 and the AGV 40. More specifically, the risk calculation unit 25 obtains a route between the worker 2 and the AGV 40 along the travelable route of the AGV 40 on the environmental map, and calculates the risk based on the distance of the route. ..
　図６Ａは、危険度算出処理の一具体例を示すフローチャートである。本例では、作業員２とＡＧＶ４０の間の距離は、単純な最短経路として求められる。 FIG. 6A is a flowchart showing a specific example of the risk calculation process. In this example, the distance between worker 2 and AGV40 is determined as a simple shortest path.
　ステップＳ１３１において、危険度算出部２５は、作業員２とＡＧＶ４０の検出位置に対応する走行可能経路のグラフ（リンク）上の位置を特定する。検出位置がリンク上にある場合は検出位置をそのまま利用すればよいが、検出位置（特に作業員２の検出位置）がリンク上に位置しないことも想定される。この場合、検出位置から最も近いリンク上の位置を利用すればよい。検出位置のリンク上での位置が特定されると、当該位置は新たなノードとしてリンクを分割する。危険度算出部２５は、分割されたリンクの距離を求める。 In step S131, the risk calculation unit 25 identifies the positions on the graph (link) of the travelable route corresponding to the detection positions of the worker 2 and the AGV 40. If the detection position is on the link, the detection position may be used as it is, but it is assumed that the detection position (particularly the detection position of the worker 2) is not located on the link. In this case, the position on the link closest to the detection position may be used. When the position of the detection position on the link is specified, the position divides the link as a new node. The risk calculation unit 25 obtains the distance of the divided links.
　なお、作業員２の検出位置とリンクとの最短距離が閾値以上である場合は、作業員２はＡＧＶ４０が走行しない安全な領域内にいると判断して、ステップＳ１３２以降の処理を行うことなく、危険度をゼロと決定してもよい。 If the shortest distance between the detection position of the worker 2 and the link is equal to or greater than the threshold value, the worker 2 determines that the AGV 40 is in a safe area where the AGV 40 does not travel, and does not perform the processing after step S132. , The risk may be determined to be zero.
　ステップＳ１３２において、危険度算出部２５は、作業員２およびＡＧＶ４０の位置の間の走行可能経路上での最短経路を求める。最短経路は、ダイクストラ法やベルマンフォード法等の公知の最短経路探索アルゴリズムによって求められる。 In step S132, the risk calculation unit 25 obtains the shortest route on the travelable route between the positions of the worker 2 and the AGV 40. The shortest path is obtained by a known shortest path search algorithm such as Dijkstra's method or Bellman-Ford method.
　ステップＳ１３３において、危険度算出部２５は、ステップＳ１３２で求められた最短経路の距離に基づいて、危険度を算出する。本実施形態では、危険度を当該距離の逆数（Risk = 1/Dist）として求める。しかしながら、危険度は距離の広義単調減少関数として求められればどのような定義式にしたがって決定されてもよい。 In step S133, the risk calculation unit 25 calculates the risk based on the distance of the shortest path obtained in step S132. In this embodiment, the degree of risk is calculated as the reciprocal of the distance (Risk = 1 / Dist). However, the degree of risk may be determined according to any definition formula as long as it is obtained as a monotonous decrease function in a broad sense of distance.
　図６Ｂ，図６Ｃを参照しつつ具体的に説明する。図６Ｂは魚眼カメラ１０によって撮影された画像（平面展開済み）の例を示す。この画像中には、１台のＡＧＶ（６０）と、２人の作業員（６１，６２）が写っており、検出部２４によって検出されている。図６Ｃは、環境地図に含まれる走行可能経路のグラフを示す。ここでは、走行可能経路のグラフは、６個のノードＮ１～Ｎ６とこれらを結ぶリンクを含む。 A specific explanation will be given with reference to FIGS. 6B and 6C. FIG. 6B shows an example of an image (planned developed) taken by the fisheye camera 10. In this image, one AGV (60) and two workers (61, 62) are shown, and they are detected by the detection unit 24. FIG. 6C shows a graph of the travelable route included in the environmental map. Here, the graph of the travelable route includes six nodes N1 to N6 and links connecting them.
　ステップＳ１３１において、ＡＧＶおよび作業員の検出位置に対応するリンク上の位置が求められる。図６Ｃのノード６３～６５がそれぞれ検出位置に対応するリンク上の位置である。以下、それぞれをノードＡ，Ｐ１，Ｐ２とも称する。これらのノードによって既存のリンクが分割されるので、分割されたリンクの距離が求められる。例えば、ノードＡとノードＮ２の距離が２，ノードＡとノードＮ５の距離が１と求められる。 In step S131, the position on the link corresponding to the detection position of the AGV and the worker is obtained. Nodes 63 to 65 in FIG. 6C are positions on the link corresponding to the detection positions, respectively. Hereinafter, each is also referred to as node A, P1 and P2. Since these nodes divide the existing link, the distance of the divided link can be obtained. For example, the distance between the node A and the node N2 is 2 and the distance between the node A and the node N5 is 1.
　ステップＳ１３２では、ＡＧＶと作業員のペアごとに、最短経路が求められる。例えば、ＡＧＶ６０と作業員６１に関しては、ノードＡ（６３）とノードＰ１（６４）の間の最短経路が求められる。ここでは、ノードＡ－ノードＮ２－ノードＮ１－ノードＰ１が、ノードＡとノードＰ１の間の最短経路として求められる。また、この経路の距離も求められる。同様に、ノードＡ（６３）とノードＰ２（６５）の間の最短経路が、ノードＡ－ノードＮ５－ノードＰ２として求められ、またその距離が求められる。 In step S132, the shortest path is obtained for each pair of AGV and worker. For example, for the AGV 60 and the worker 61, the shortest path between the node A (63) and the node P1 (64) is required. Here, node A-node N2-node N1-node P1 is obtained as the shortest path between node A and node P1. The distance of this route is also required. Similarly, the shortest path between node A (63) and node P2 (65) is determined as node A-node N5-node P2, and its distance is determined.
　ステップＳ１３３では、危険度Riskが、作業員とＡＧＶの間の経路の距離Distに基づいてRisk = 1/Distと決定される。ここでは、ＡＧＶ６０と作業員６１の間の危険度は、Risk=1/3.5=0.29と算出され、ＡＧＶ６０と作業員６２の間の危険度はRisk=1/1.5=0.67と算出される。ＡＧＶ６０と作業員６１，６２の間の直線距離はほぼ同一であるが、走行経路を考慮することで、ＡＧＶ６０と作業員６２の間の危険度を、ＡＧＶ６０と作業員６１の間の危険度よりも大きく算出することができる。 In step S133, the risk level Risk is determined as Risk = 1 / Dist based on the distance Dist of the route between the worker and the AGV. Here, the risk between the AGV 60 and the worker 61 is calculated as Risk = 1 / 3.5 = 0.29, and the risk between the AGV 60 and the worker 62 is calculated as Risk = 1 / 1.5 = 0.67. The straight-line distance between the AGV 60 and the workers 61 and 62 is almost the same, but by considering the traveling route, the risk between the AGV 60 and the worker 62 can be determined from the risk between the AGV 60 and the worker 61. Can also be calculated large.
　図７Ａは、危険度算出処理の他の具体例を示すフローチャートである。本例では、作業員とＡＧＶの移動速度を考慮して、作業員とＡＧＶの間の経路を求める。なお、本例においては、作業員およびＡＧＶの移動速度（方向および速さ）が求められているものとする。移動速度は、直近の複数フレームにおける作業員およびＡＧＶの位置から求めることができる。図７Ｂに、ＡＧＶ６０および作業員６１，６２の位置とその移動速度が表されている（矢印７０～７２）。 FIG. 7A is a flowchart showing another specific example of the risk calculation process. In this example, the path between the worker and the AGV is determined in consideration of the moving speed of the worker and the AGV. In this example, it is assumed that the moving speed (direction and speed) of the worker and the AGV is required. The moving speed can be determined from the positions of the worker and the AGV in the most recent multiple frames. FIG. 7B shows the positions of the AGV 60 and the workers 61 and 62 and their moving speeds (arrows 70 to 72).
　ステップＳ１３１において、危険度算出部２５は、作業員とＡＧＶの検出位置に対応する走行可能経路のグラフ（リンク）上の位置を特定する。この処理は上記と同様であるので説明を省略する。 In step S131, the risk calculation unit 25 specifies a position on the graph (link) of the travelable route corresponding to the detection position of the worker and the AGV. Since this process is the same as above, the description thereof will be omitted.
　ステップＳ１３５において、危険度算出部２５は、作業員とＡＧＶのうち速さが速い方を選択する。この選択は、作業員とＡＧＶのペアごとに行われる。例えば、図７Ｂに示す状況では、ＡＧＶ６０と作業員６１に関しては作業員６１の方が選択され、ＡＧＶ６０と作業員６２に関してはＡＧＶ６０が選択される。 In step S135, the risk calculation unit 25 selects the worker and the AGV, whichever is faster. This selection is made for each worker-AGV pair. For example, in the situation shown in FIG. 7B, the worker 61 is selected for the AGV 60 and the worker 61, and the AGV 60 is selected for the AGV 60 and the worker 62.
　ステップＳ１３６において、危険度算出部２５は、ステップＳ１３５で選択された作業員またはＡＧＶ（以下、高速移動物体と称する）が、次に到達するノードを特定する。ステップＳ１３７において、危険度算出部２５は、ステップＳ１３６で特定されたノードと、ステップＳ１３５で選択されなかった作業員またはＡＧＶ（以下、低速移動物体と称する）との間の、最短経路を求める。ステップＳ１３８において、危険度算出部２５は、高速移動物体の位置と次ノードを結ぶ経路と、ステップＳ１３５で求められた最短経路を足し合わせた経路を、作業員とＡＧＶの間の経路として決定する。すなわち、本例のステップＳ１３５～Ｓ１３８の処理では、走行可能経路に沿った作業員とＡＧＶの間の経路のうち、高速移動物体の進行方向にあるノードを経由する最短経路が決定される。 In step S136, the risk calculation unit 25 identifies the node to which the worker or AGV (hereinafter referred to as a high-speed moving object) selected in step S135 will reach next. In step S137, the risk calculation unit 25 finds the shortest path between the node identified in step S136 and the worker or AGV (hereinafter referred to as a low-speed moving object) not selected in step S135. In step S138, the risk calculation unit 25 determines as a route between the worker and the AGV, which is the sum of the route connecting the position of the high-speed moving object and the next node and the shortest route obtained in step S135. .. That is, in the processing of steps S135 to S138 of this example, among the routes between the worker and the AGV along the travelable route, the shortest route via the node in the traveling direction of the high-speed moving object is determined.
　図７Ｃを参照して、ＡＧＶ６０と作業員６１の組み合わせについて具体的に説明する。この例では、作業員６１の方が速く移動しているので、作業員６１が高速移動物体として選択され、ＡＧＶ６０が低速移動物体として選択される（Ｓ１３５）。作業員６１は、図中右方向に移動しているので、次に到達するノードはノードＮ４である（Ｓ１３６）。したがって、ノードＮ４とノードＡ（７３）の最短経路として経路７５が求められ（Ｓ１３７）、ノードＰ１（７４）とノードＮ４を結ぶ経路７６と経路７５を足し合わせた経路が、作業員６１とＡＧＶ６０の間の経路として決定される（Ｓ１３８）。 The combination of the AGV 60 and the worker 61 will be specifically described with reference to FIG. 7C. In this example, since the worker 61 is moving faster, the worker 61 is selected as the high-speed moving object and the AGV60 is selected as the low-speed moving object (S135). Since the worker 61 is moving to the right in the figure, the next node to be reached is node N4 (S136). Therefore, the route 75 is obtained as the shortest route between the node N4 and the node A (73) (S137), and the route connecting the node P1 (74) and the node N4 and the route 75 are added together to form the worker 61 and the AGV60. It is determined as a route between (S138).
　図７Ｄを参照して、ＡＧＶ６０と作業員６２の組み合わせについて具体的に説明する。この例では、ＡＧＶ６０の方が速く移動しているので、ＡＧＶ６０が高速移動物体として選択され、作業員６２が低速移動物体として選択される（Ｓ１３５）。ＡＧＶ６０は、図中右方向に移動しているので、次に到達するノードはノードＮ５である（Ｓ１３６）。したがって、ノードＮ５とノードＰ２（７７）の最短経路として経路７８が求められ（Ｓ１３７）、ノードＡ（７３）とノードＮ５を結ぶ経路７９と経路７９を足し合わせた経路が、作業員６２とＡＧＶ６０の間の経路として決定される（Ｓ１３８）。 The combination of the AGV 60 and the worker 62 will be specifically described with reference to FIG. 7D. In this example, since the AGV60 is moving faster, the AGV60 is selected as the high-speed moving object and the worker 62 is selected as the low-speed moving object (S135). Since the AGV60 is moving to the right in the figure, the next node to reach is node N5 (S136). Therefore, the route 78 is obtained as the shortest route between the node N5 and the node P2 (77) (S137), and the route 79 and the route 79 connecting the node A (73) and the node N5 are added to the worker 62 and the AGV60. It is determined as a route between (S138).
　ステップＳ１３９において、危険度算出部２５は、作業員とＡＧＶの間の経路の距離および作業員とＡＧＶの移動速度に基づいて危険度を算出する。ここで、作業員とＡＧＶの間の距離をDist, 作業員の速さをVp, ＡＧＶの速さをVAGPとすると、危険度Riskは、例えば、
　　Risk = Max(Vp, VAGP)/Dist
　によって求めることができる。 In step S139, the risk calculation unit 25 calculates the risk based on the distance of the route between the worker and the AGV and the moving speed of the worker and the AGV. Here, assuming that the distance between the worker and the AGV is Dist, the speed of the worker is V p , and the speed of the AGV is V AGP , the risk risk is, for example,
Risk = Max (V p , V AGP ) / Dist
Can be obtained by.
　この例では、危険度は距離に反比例しているが、距離が小さいほど危険度が大きく算出されればよい。また、この例では、作業員の速さとＡＧＶの速さのうちの速い方の値（最大値）が大きいほど危険度が大きく算出されるが、他の方法で危険度を求めてもよい。例えば、作業員とＡＧＶの相対速度が大きいほど危険度を大きく算出してもよい。ここで相対速度は、作業員とＡＧＶを結ぶ経路の方向を考慮して算出され、異なる方向（向かい合う方向）に進んでいる場合は速さの和が相対速度となり、同じ方向に進んでいる場合は速さの差（絶対値）が相対速度となる。 In this example, the degree of risk is inversely proportional to the distance, but the smaller the distance, the greater the degree of risk should be calculated. Further, in this example, the larger the value (maximum value) of the speed of the worker and the speed of the AGV is, the larger the risk is calculated, but the risk may be obtained by another method. For example, the greater the relative speed between the worker and the AGV, the greater the risk may be calculated. Here, the relative speed is calculated in consideration of the direction of the path connecting the worker and the AGV, and the sum of the speeds is the relative speed when traveling in different directions (opposing directions), and when traveling in the same direction. The difference in speed (absolute value) is the relative speed.
　なお、本例においても、速度を考慮せずに経路の距離のみに基づいて危険度を算出しても構わない。 In this example as well, the degree of risk may be calculated based only on the distance of the route without considering the speed.
　図３に戻って、ステップＳ１４のＡＧＶ制御指令の作成・送信処理について説明する。 Returning to FIG. 3, the process of creating and transmitting the AGV control command in step S14 will be described.
　図８Ａは、ステップＳ１４のＡＧＶ制御指令の生成・送信処理の詳細を示すフローチャートである。制御指令送信部２６は、ステップＳ１３で求めた危険度Ｒｉｓｋを閾値ＴＨ１，ＴＨ２（ＴＨ１＜ＴＨ２）と比較して、比較結果に基づいてＡＧＶ４０の走行速度を決定する。 FIG. 8A is a flowchart showing details of the AGV control command generation / transmission process in step S14. The control command transmission unit 26 compares the risk risk obtained in step S13 with the threshold values TH1 and TH2 (TH1 <TH2), and determines the traveling speed of the AGV40 based on the comparison result.
　制御指令送信部２６は、危険度Ｒｉｓｋが第１の閾値ＴＨ１以下の場合（Ｓ１４１－ＹＥＳ）は走行速度をＶ１に決定（Ｓ１４２）し、危険度Ｒｉｓｋが第２の閾値ＴＨ２以下の場合（Ｓ１４３－ＹＥＳ）は走行速度をＶ２（Ｖ２＜Ｖ１）に決定し（Ｓ１４４）、それ以外の場合（Ｓ１４３－ＮＯ）は走行速度をゼロに決定する（Ｓ１４５）。そして、制御指令送信部２６は、上記で決定した走行速度を含む制御指令を生成して、ホスト装置３０に送信する（Ｓ１４６）。 The control command transmission unit 26 determines the traveling speed to V1 (S142) when the risk level Risk is equal to or less than the first threshold value TH1 (S141-YES), and determines the traveling speed to V1 (S142), and when the risk level Risk is equal to or less than the second threshold value TH2 (S143). -YES) determines the traveling speed to V2 (V2 <V1) (S144), otherwise (S143-NO) determines the traveling speed to zero (S145). Then, the control command transmission unit 26 generates a control command including the traveling speed determined above and transmits it to the host device 30 (S146).
　図８Ｂは、上記のフローチャートに従う制御における、危険度ＲｉｓｋとＡＧＶの走行速度Ｖの関係を示すグラフである。ここでは、２つの閾値によって危険度を３つのレベルに分けて、レベル毎に予め定められた速さ（Ｖ１，Ｖ２，０）でＡＧＶが移動するように制御している。しかしながら、レベルの数は３つである必要は無く、２レベルあるいは４レベル以上に分けても構わない。また、この例では、速度を上げる場合と下げる場合で同じ閾値を採用しているが、速度を上げる場合と下げる場合で異なる閾値を採用してヒステリシスを持たせてもよい。 FIG. 8B is a graph showing the relationship between the risk risk and the traveling speed V of the AGV in the control according to the above flowchart. Here, the risk level is divided into three levels according to two threshold values, and the AGV is controlled to move at a predetermined speed (V1, V2, 0) for each level. However, the number of levels does not have to be three and may be divided into two levels or four or more levels. Further, in this example, the same threshold value is adopted when the speed is increased and when the speed is decreased, but different threshold values may be adopted when the speed is increased and decreased to have hysteresis.
　また、図８Ａのフローチャートでは、処理ごとにＡＧＶの制御指令を生成・送信するように示しているが、ＡＧＶの走行速度を変化させる必要が生じた場合のみ、制御指令を生成・送信するようにしても構わない。 Further, in the flowchart of FIG. 8A, it is shown that the control command of the AGV is generated and transmitted for each process, but the control command is generated and transmitted only when it becomes necessary to change the traveling speed of the AGV. It doesn't matter.
　以上のように、本実施形態によれば、作業員とＡＧＶの間の単純な直線距離ではなく、ＡＧＶの移動可能経路に沿った経路の距離に基づいて危険度を算出しているので、作業員とＡＧＶが接触する危険性をより適切に評価できる。したがって、本来は接触の危険性が低いにもかかわらずＡＧＶを停止させてしまうような事態を避けることができる。 As described above, according to the present embodiment, the risk level is calculated based on the distance of the path along the movable path of the AGV, not on the simple linear distance between the worker and the AGV. The risk of contact between personnel and AGV can be evaluated more appropriately. Therefore, it is possible to avoid a situation in which the AGV is stopped even though the risk of contact is originally low.
　＜その他＞
　上記実施形態は、本発明の構成例を例示的に説明するものに過ぎない。本発明は上記の具体的な形態には限定されることはなく、その技術的思想の範囲内で種々の変形が可能である。 <Others>
The above-described embodiment merely exemplifies a configuration example of the present invention. The present invention is not limited to the above-mentioned specific form, and various modifications can be made within the scope of its technical idea.
　上記の実施形態において、作業員およびＡＧＶの位置を魚眼カメラによる撮影画像から検出しているが、魚眼カメラ以外のカメラによる撮影画像から検出してもよい。また、画像以外から作業員およびＡＧＶの位置を検出してもよい。例えば、作業員およびＡＧＶにビーコン（電波）を発信するビーコン発信器を取り付け、ＡＧＶ制御装置２０がビーコン受信器によってビーコンを受信することで、作業員およびＡＧＶの位置を検出してもよい。なお、電波以外に、超音波や可視光を用いて位置を検出してもよい。 In the above embodiment, the positions of the worker and the AGV are detected from the image taken by the fisheye camera, but it may be detected from the image taken by a camera other than the fisheye camera. Further, the positions of the worker and the AGV may be detected from other than the image. For example, a beacon transmitter that transmits a beacon (radio wave) may be attached to the worker and the AGV, and the AGV control device 20 may detect the position of the worker and the AGV by receiving the beacon by the beacon receiver. In addition to radio waves, the position may be detected using ultrasonic waves or visible light.
　上記の実施形態において、ＡＧＶ制御装置２０が危険度に基づいて制御指令を生成してホスト装置３０に送信し、ホスト装置３０からＡＧＶ４０に制御指令を送信している。しかしながら、ＡＧＶ制御装置２０は危険度のみをホスト装置３０に送信して、ホスト装置３０がＡＧＶの制御を決定してもよい。また、ＡＧＶ制御装置２０がＡＧＶ４０に対して制御指令を直接送信してもよい。 In the above embodiment, the AGV control device 20 generates a control command based on the degree of risk and transmits it to the host device 30, and the host device 30 transmits the control command to the AGV 40. However, the AGV control device 20 may transmit only the degree of risk to the host device 30, and the host device 30 may determine the control of the AGV. Further, the AGV control device 20 may directly transmit a control command to the AGV 40.
　＜付記＞
　１．作業員（２）および無人搬送車（４０）が移動する環境における前記作業員（２）および前記無人搬送車（４０）の位置を検出する検出手段（２４）と、
　前記環境において前記無人搬送車（４０）の走行可能経路を含む環境地図を参照して、前記走行可能経路に沿った前記作業員（２）と前記無人搬送車（４０）の間の距離に基づいて、前記無人搬送車（４０）が前記作業員（２）に与える危険度を算出する算出手段（２５）と、
　を備える、危険度算出装置（２０）。 <Additional notes>
1. 1. A detection means (24) for detecting the positions of the worker (2) and the automatic guided vehicle (40) in an environment in which the worker (2) and the automatic guided vehicle (40) move.
Based on the distance between the worker (2) and the automatic guided vehicle (40) along the travelable route with reference to an environmental map including the travelable route of the automatic guided vehicle (40) in the environment. Then, the calculation means (25) for calculating the degree of danger given to the worker (2) by the automatic guided vehicle (40), and
A risk calculation device (20).
　２．コンピュータ（２０）によって実行される危険度算出方法であって、
　作業員（２）および無人搬送車（４０）が移動する環境における前記作業員（２）および前記無人搬送車（４０）の位置を検出する検出ステップ（Ｓ１２）と、
　前記環境において前記無人搬送車（４０）の走行可能経路を含む環境地図を参照して、前記走行可能経路に沿った前記作業員（２）と前記無人搬送車（４０）の間の距離に基づいて、前記無人搬送車（４０）が前記作業員（２）に与える危険度を算出する算出ステップ（Ｓ１３）と、
　を含む、危険度算出方法。 2. A risk calculation method performed by the computer (20).
A detection step (S12) for detecting the positions of the worker (2) and the automatic guided vehicle (40) in an environment in which the worker (2) and the automatic guided vehicle (40) move.
Based on the distance between the worker (2) and the automatic guided vehicle (40) along the travelable route with reference to an environmental map including the travelable route of the automatic guided vehicle (40) in the environment. The calculation step (S13) for calculating the degree of risk given to the worker (2) by the automatic guided vehicle (40) and
Risk calculation method including.
１０：魚眼カメラ　　２０：ＡＧＶ制御装置　　３０：ホスト装置　　４０：ＡＧＶ
２１：画像取得部　　２２：環境地図作成部　　２３：地図記憶部
２４：検出部　　　　２５：危険度算出部　　　２６：制御指令送信部 10: Fisheye camera 20: AGV control device 30: Host device 40: AGV
21: Image acquisition unit 22: Environmental map creation unit 23: Map storage unit 24: Detection unit 25: Risk calculation unit 26: Control command transmission unit

Claims (17)

  1. 　作業員および無人搬送車が移動する環境における前記作業員および前記無人搬送車の位置を検出する検出手段と、
    　前記環境において前記無人搬送車の走行可能経路を含む環境地図を参照して、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の距離に基づいて、前記無人搬送車が前記作業員に与える危険度を算出する算出手段と、
    　を備える、危険度算出装置。     A detection means for detecting the positions of the worker and the automatic guided vehicle in an environment in which the worker and the automatic guided vehicle move, and
    With reference to an environmental map including a travelable route of the automatic guided vehicle in the environment, the automatic guided vehicle performs the work based on the distance between the worker and the automatic guided vehicle along the travelable route. A calculation method for calculating the degree of risk given to members,
    A risk calculation device equipped with.
  2. 　前記算出手段は、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の最短経路の距離に基づいて、前記危険度を算出する、
    　請求項１に記載の危険度算出装置。     The calculation means calculates the degree of risk based on the distance of the shortest path between the worker and the automatic guided vehicle along the travelable route.
    The risk calculation device according to claim 1.
  3. 　前記算出手段は、前記作業員と前記無人搬送車のうち速さが速い方の進行方向にあるノードを経由する、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の最短経路の距離に基づいて、前記危険度を算出する、
    　請求項１に記載の危険度算出装置。     The calculation means is the shortest route between the worker and the automatic guided vehicle along the travelable route via a node in the traveling direction of the worker and the automatic guided vehicle, whichever has the higher speed. Calculate the degree of risk based on the distance of
    The risk calculation device according to claim 1.
  4. 　前記環境地図は、前記走行可能経路がグラフ形式で表されており、
    　前記算出手段は、最短経路探索アルゴリズムを用いて、前記最短経路を求める、
    　請求項２または３に記載の危険度算出装置。     In the environmental map, the travelable route is represented in a graph format.
    The calculation means obtains the shortest path by using the shortest path search algorithm.
    The risk calculation device according to claim 2 or 3.
  5. 　前記算出手段は、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の距離が小さいほど、前記危険度を大きく算出する、
    　請求項１から４のいずれか１項に記載の危険度算出装置。     The calculation means calculates the degree of risk as the distance between the worker and the automatic guided vehicle along the travelable route is smaller.
    The risk calculation device according to any one of claims 1 to 4.
  6. 　前記算出手段は、前記作業員の移動の速さおよび前記無人搬送車の移動の速さの少なくとも一方も考慮して、前記危険度を算出する、
    　請求項１から５のいずれか１項に記載の危険度算出装置。     The calculation means calculates the degree of risk in consideration of at least one of the moving speed of the worker and the moving speed of the automatic guided vehicle.
    The risk calculation device according to any one of claims 1 to 5.
  7. 　前記算出手段は、前記作業員の速さと前記無人搬送車の速さのうちの最大値が大きいほど、前記危険度を大きく算出する、
    　請求項６に記載の危険度算出装置。     The calculation means calculates the degree of risk as the maximum value of the speed of the worker and the speed of the automatic guided vehicle increases.
    The risk calculation device according to claim 6.
  8. 　前記算出手段は、前記作業員と前記無人搬送車の相対速度が大きいほど、前記危険度を大きく算出する、
    　請求項６に記載の危険度算出装置。     The calculation means calculates the degree of risk as the relative speed between the worker and the automatic guided vehicle increases.
    The risk calculation device according to claim 6.
  9. 　前記環境の画像を撮影するカメラをさらに備え、
    　前記検出手段は、前記画像に対する画像処理によって、前記作業員および前記無人搬送車の位置を検出する、
    　請求項１から８のいずれか１項に記載の危険度算出装置。     Further equipped with a camera that captures an image of the environment
    The detection means detects the positions of the worker and the automatic guided vehicle by image processing the image.
    The risk calculation device according to any one of claims 1 to 8.
  10. 　前記作業員および前記無人搬送車が有するビーコン発信器から送信される電波を受信する受信器をさらに備え、
    　前記検出手段は、前記受信器の受信結果に基づいて、前記作業員および前記無人搬送車の位置を検出する、
    　請求項１から８のいずれか１項に記載の危険度算出装置。     Further provided with a receiver for receiving radio waves transmitted from the beacon transmitter included in the worker and the automatic guided vehicle.
    The detection means detects the positions of the worker and the automatic guided vehicle based on the reception result of the receiver.
    The risk calculation device according to any one of claims 1 to 8.
  11. 　前記環境を撮影した画像から前記無人搬送車を検出し、当該検出の結果に基づいて前記環境地図を作成する地図作成手段をさらに備える、
    　請求項１から１０のいずれか１項に記載の危険度算出装置。     A map creation means for detecting the automatic guided vehicle from an image of the environment and creating the environment map based on the result of the detection is further provided.
    The risk calculation device according to any one of claims 1 to 10.
  12. 　位置検出手段を備えた無人搬送車から送信される位置情報に基づいて、前記環境地図を作成する地図作成手段をさらに備える、
    　請求項１から１０のいずれか１項に記載の危険度算出装置。     A cartographic means for creating the environmental map based on the position information transmitted from the automatic guided vehicle provided with the position detecting means is further provided.
    The risk calculation device according to any one of claims 1 to 10.
  13. 　前記環境を撮影した画像におけるユーザの設定に基づいて、前記環境地図を作成する地図作成手段をさらに備える、
    　請求項１から１０のいずれか１項に記載の危険度算出装置。     A map creation means for creating the environment map based on the user's setting in the image obtained by capturing the environment is further provided.
    The risk calculation device according to any one of claims 1 to 10.
  14. 　請求項１から１３のいずれか１項に記載の危険度算出装置と、
    　前記算出手段によって算出された前記危険度に応じて、前記無人搬送車の移動速度を制御する制御手段と、
    　を備える制御装置。     The risk calculation device according to any one of claims 1 to 13.
    A control means for controlling the moving speed of the automatic guided vehicle according to the degree of danger calculated by the calculation means, and
    A control device comprising.
  15. 　前記制御手段は、前記危険度をレベル分けし、危険度のレベルに応じた速度で前記無人搬送車が移動するように制御する、
    　請求項１４に記載の制御装置。     The control means divides the risk level into levels and controls the automatic guided vehicle to move at a speed according to the risk level.
    The control device according to claim 14.
  16. 　コンピュータによって実行される危険度算出方法であって、
    　作業員および無人搬送車が移動する環境における前記作業員および前記無人搬送車の位置を検出する検出ステップと、
    　前記環境において前記無人搬送車の走行可能経路を含む環境地図を参照して、前記走行可能経路に沿った前記作業員と前記無人搬送車の間の距離に基づいて、前記無人搬送車が前記作業員に与える危険度を算出する算出ステップと、
    　を含む、危険度算出方法。     A risk calculation method performed by a computer
    A detection step for detecting the positions of the worker and the automatic guided vehicle in an environment in which the worker and the automatic guided vehicle move, and a detection step.
    With reference to an environmental map including a travelable route of the automatic guided vehicle in the environment, the automatic guided vehicle performs the work based on the distance between the worker and the automatic guided vehicle along the travelable route. Calculation steps to calculate the degree of risk given to members, and
    Risk calculation method including.
  17. 　コンピュータを請求項１から１５のいずれか１項に記載の装置の各手段として機能させるためのプログラム。 A program for making a computer function as each means of the device according to any one of claims 1 to 15.
PCT/JP2020/047022 2020-03-09 2020-12-16 Danger degree calculation device, and device and method for controlling unmanned carrier vehicle WO2021181791A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2020-039704 2020-03-09
JP2020039704A JP2021140638A (en) 2020-03-09 2020-03-09 Danger degree calculation device, control device for unmanned dolly, and method therefor

Publications (1)

Publication Number Publication Date
WO2021181791A1 true WO2021181791A1 (en) 2021-09-16

Family

ID=77668792

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2020/047022 WO2021181791A1 (en) 2020-03-09 2020-12-16 Danger degree calculation device, and device and method for controlling unmanned carrier vehicle

Country Status (2)

Country Link
JP (1) JP2021140638A (en)
WO (1) WO2021181791A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012038011A (en) * 2010-08-05 2012-02-23 Ihi Corp Route generation apparatus for mobile object
WO2018062336A1 (en) * 2016-09-30 2018-04-05 日本電気株式会社 Flight control device, unmanned aerial vehicle, flight control method, and computer-readable recording medium
JP2019109879A (en) * 2017-12-18 2019-07-04 ザ・ボーイング・カンパニーＴｈｅ Ｂｏｅｉｎｇ Ｃｏｍｐａｎｙ Multisensor safety route system for autonomous vehicles

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012038011A (en) * 2010-08-05 2012-02-23 Ihi Corp Route generation apparatus for mobile object
WO2018062336A1 (en) * 2016-09-30 2018-04-05 日本電気株式会社 Flight control device, unmanned aerial vehicle, flight control method, and computer-readable recording medium
JP2019109879A (en) * 2017-12-18 2019-07-04 ザ・ボーイング・カンパニーＴｈｅ Ｂｏｅｉｎｇ Ｃｏｍｐａｎｙ Multisensor safety route system for autonomous vehicles

Also Published As

Publication number Publication date
JP2021140638A (en) 2021-09-16

Similar Documents

Publication Publication Date Title
US11340610B2 (en) Autonomous target following method and device
JP7341652B2 (en) Information processing device, information processing method, program, and system
US11673269B2 (en) Method of identifying dynamic obstacle and robot implementing same
US8090193B2 (en) Mobile robot
US7873448B2 (en) Robot navigation system avoiding obstacles and setting areas as movable according to circular distance from points on surface of obstacles
US20100222925A1 (en) Robot control apparatus
JP6771588B2 (en) Moving body and control method of moving body
US11822334B2 (en) Information processing apparatus, information processing method, and program for control of a moving body capable of autonomous movement
KR102560462B1 (en) Moving robot
JPWO2019187816A1 (en) Mobiles and mobile systems
US20220291685A1 (en) Method and system to improve autonomous robotic systems responsive behavior
JPWO2018180454A1 (en) Moving body
TWI532619B (en) Dual Image Obstacle Avoidance Path Planning Navigation Control Method
Bista et al. Image-based indoor topological navigation with collision avoidance for resource-constrained mobile robots
US11906970B2 (en) Information processing device and information processing method
WO2021181791A1 (en) Danger degree calculation device, and device and method for controlling unmanned carrier vehicle
KR20210042537A (en) Method of estimating position in local area in large sapce and robot and cloud server implementing thereof
US20190354102A1 (en) Autonomous robotic system
Yuan et al. Visual steering of UAV in unknown environments
JP2009223632A (en) Autonomous mobile device and autonomous mobile device system
Jeon et al. Indoor/outdoor transition recognition based on door detection
JP2022012173A (en) Information processing device, information processing system, information processing method, and program
Zhang et al. Autonomous indoor exploration of mobile robots based on door-guidance and improved dynamic window approach
US20230297121A1 (en) Moving body control system, control apparatus, and moving body control method
Suprapto et al. Designing an Autonomous Vehicle Using Sensor Fusion Based on Path Planning and Deep Learning Algorithms

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20924453

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20924453

Country of ref document: EP

Kind code of ref document: A1