Unmanned carrier and unmanned-carrier system

Abstract

To provide an unmanned carrier and unmanned-carrier system capable of identifying a coupling point of a transport cart so as to appropriately couple to the transport cart even if no marker is put on the transport cart.SOLUTION: An unmanned carrier 10 for automatically carrying a transport cart 40 to a destination comprises a coupling device for coupling to the transport cart 40, a camera that takes an image of the transport cart 40, a range-finding sensor that emits and receives a laser beam of light and measures a distance to and shape of the transport cart, and a control unit. The control unit has a function for recognizing a shape of the transport cart concerned based on a taken image of the camera, a function for identifying a coupling point of the transport cart concerned based on measurement data of the range-finding sensor, a function for moving the unmanned carrier 10 such that the coupling point and the coupling device become to have a desired relative angle, a function for coupling the coupling device to the coupling point and carrying the transport cart 40 to a destination, and a function for releasing the coupling device from coupling at the destination.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an automated guided vehicle and an automated guided vehicle system that automatically connect to a transport cart such as a cage cart and transport the transport cart to a destination.

Background Art

[0002] Conventionally, an automated guided vehicle that automatically connects to a cage cart and transports the cage cart to a destination is known. For example, in Patent Document 1, a marker (for example, a QR code (registered trademark)) is attached to a cage cart, and an automated guided vehicle recognizes the marker to identify the connection location of the cage cart and connects to the identified connection location.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in Patent Document 1, it is necessary to attach a marker to each cage cart, and after transporting the cage cart, it is necessary to remove the marker from the cage cart and attach the marker to a newly transported cage cart. As the number of cage carts increases, the labor of workers increases.

[0005] An object of the present invention is to provide an automated guided vehicle and an automated guided vehicle system that can identify the connection location of a transport cart and appropriately connect to the transport cart without attaching a marker to the transport cart such as a cage cart.

Means for Solving the Problems

[0006] The automated guided vehicle according to the present invention is an automated guided vehicle that automatically transports a transport cart to a destination, and includes a coupling device for coupling to the transport cart, a camera for imaging the transport cart, a distance measuring sensor that irradiates and receives laser light to measure the distance and shape of the transport cart, and a control unit. The control unit has a cart shape recognition function for recognizing the shape of the target transport cart based on the captured image of the camera, a connection position specifying function for specifying the connection location of the target transport cart based on the measurement data of the distance measuring sensor, a relative position adjustment function for moving the automated guided vehicle so that the connection location and the coupling device have a desired relative angle, a connection and unloading function for connecting the coupling device to the connection location and transporting the transport cart to the destination, and a release function for releasing the connection of the coupling device at the destination. In the above automated guided vehicle, the control unit stores map information including a hooking area, and may further include a target cart detection function for detecting whether or not the target transport cart exists in the hooking area based on the measurement result of the distance measuring sensor. In the above automated guided vehicle, when the target cart detection function of the control unit hooks a transport cart arranged in a predetermined hooking area, the hooking area may be divided into a plurality of lanes, and one of the lanes in which the transport cart exists may be set as a target lane in which the target transport cart exists. In the above automated guided vehicle, the cart shape recognition function of the control unit has a function of calculating a three-dimensional rectangular area of the transport cart by performing image recognition processing on the image captured by the camera, and the relative position adjustment function of the control unit is based on the calculated three-dimensional rectangular area and the measurement data of the distance measuring sensor. By specifying the connection surface of the transport cart and the connection location on the connection surface, the relative angle of the automated guided vehicle with respect to the transport cart can be adjusted. In the above automated guided vehicle, the image captured by the camera may be a three-dimensional image, and the measurement data of the distance measuring sensor may be point cloud data. In the above-described driverless transport vehicle, the camera can be a stereo camera, and the distance measuring sensor can be an omnidirectional LiDAR. In the above-described driverless transport vehicle, the control unit stores map information including a release area, and when the release function of the control unit releases the transport cart to a predetermined release area, the release area is divided into a plurality of lanes, and one of the lanes where the transport cart does not exist is set as the destination. In the above-described driverless transport vehicle, the coupling device can be configured to include a hook that engages with the coupling location, a contact surface that contacts the coupling location, and a driving device that drives the hook. In the above-described driverless transport vehicle, further provided is a wireless communication device that receives control information including the vehicle type information of the transport cart and the coupling location information for each vehicle type from an external server. The cart shape recognition function recognizes the shape of the target transport cart based on the control information and the captured image of the camera, and the coupling position specifying function specifies the coupling location of the target transport cart based on the control information and the distance measurement result of the distance measuring sensor. The server according to the present invention is a server configured to be communicable with the above-described driverless transport vehicle, and transmits the vehicle type information of the transport cart used in the cart shape recognition function and the coupling location information for each vehicle type of the transport cart used in the coupling position specifying function to the driverless transport vehicle. The management terminal according to the present invention is a management terminal configured to be communicable with the above-described server, and transmits the vehicle type information of the transport cart used in the cart shape recognition function and the coupling location information for each vehicle type of the transport cart used in the coupling position specifying function to the server.

Advantages of the Invention

[0007] According to the present invention, even without attaching a marker to the transport cart, the coupling location of the transport cart can be appropriately specified, and the coupling part can be appropriately coupled to the specified coupling location.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Embodiments for Carrying Out the Invention

[0009] Hereinafter, with reference to the drawings, embodiments of the automated guided vehicle system according to the present invention will be described. As shown in FIG. 1, the automated guided vehicle system 1 according to this embodiment includes an RM server 30, a plurality of carrier vehicles 40, a management terminal 50, and a plurality of automated guided vehicles 10. The carrier vehicles 40 and the automated guided vehicles 10 are installed at workplaces in indoor facilities such as warehouses. In this embodiment, the carrier vehicle 40 is exemplified and described as the transport trolley transported by the automated guided vehicle 10. However, the trolley transported by the automated guided vehicle 10 is not limited to the carrier vehicle 40, and transport trolleys such as flatbed trolleys and cart trucks can also be used.

[0010] The RM server 30 is a server capable of realizing the management of the automated guided vehicle 10, and is constructed by installing RM (Robot Master) software and database software in a server device equipped with a CPU and a storage device. This RM software includes, as software modules for realizing main functions, a carrier vehicle management unit 31, an operation adjustment unit 32, and a workplace management unit 33.

[0011] The carrier vehicle management unit 31 manages the workplace information where the automated guided vehicle 10 is located and the operation information of the automated guided vehicle 10. In the carrier vehicle management unit 31, the vehicle ID of the automated guided vehicle 10 is associated with the workplace ID of the work target and managed. The operation adjustment unit 32 manages the operation time of a plurality of automated guided vehicles 10. Note that there may be cases where a plurality of automated guided vehicles 10 operate in the same space within the same building, or there may be cases where only one operates. The workplace management unit 33 manages the basic map for each workplace and transmits the basic map to the automated guided vehicle 10. In the basic map, a passable course and a prohibited area are set. Note that the management of the basic map of the workplace may not be performed by the RM server 30, and may be configured to be managed by the automated guided vehicle 10 and the management terminal 50 installed at the workplace.

[0012] Also, in the present embodiment, the RM server 30 has a conveyance operation control unit 34 for controlling the conveyance operation in which the automated guided vehicle 10 conveys the cage cart 40. The conveyance operation control unit 24 stores the patrol course of the automated guided vehicle 10, the type information of the cage cart 40 to be conveyed by the automated guided vehicle 10, and the control information of the connection points for each type, based on the input from the administrator or operator via the management terminal 50, and transmits this control information to the automated guided vehicle 10.

[0013] The management terminal 50 is a terminal such as a notebook PC, a desktop PC, or a tablet, and is used for managing the RM server 30. The administrator or operator can manage a plurality of automated guided vehicles 10 by accessing the RM server 30 using the management terminal 50. Also, as described above, the administrator or operator can register the control information of the automated guided vehicle 10 in the RM server 30 using the management terminal 50.

[0014] The cage cart 40 is an example of a conveyance trolley for loading and transporting goods. The cage cart 40 has wheels and can transport the loaded goods by being towed by the automated guided vehicle 10. Here, FIG. 2 is a perspective view showing an example of the cage cart 40. As shown in FIG. 2, the cage cart 40 according to the present embodiment has a rectangular parallelepiped or cubic shape, and at least one of the side surfaces that are planes is a connection surface 41 to which the automated guided vehicle 10 is connected. The connection surface 41 is composed of a lattice of poles 42 extending in the vertical direction and bars 43 extending in the horizontal direction. A bar 43 installed at a height close to the height position of the hook 14a of the connection device 14 between the poles 42 of the connection surface 41 is a connection point 44 to which the connection device 14 of the automated guided vehicle 10 is connected. The connection point 44 is a location specified by the automated guided vehicle 10. For example, in the example shown in FIG. 4, the automated guided vehicle 10 can set the vicinity of the center of a bar 43a installed at a height close to the height position of the hook 14a of the connection device 14 between the poles 42a and 42b of the connection surface 41 as the connection point 44 to which the connection device 14 of the automated guided vehicle 10 is connected.

[0015] Based on the control of the RM server 30, the automated guided vehicle 10 has the function of automatically connecting to the cart 40 placed in the predetermined hooking area F in the warehouse and transporting the connected cart 40 to the release area R. In particular, in this embodiment, the automated guided vehicle 10 is characterized in that it can identify the connection point 44 of the cart 40 and properly connect to the cart 40 without attaching a marker such as a QR code (registered trademark) to the cart 40. The details of the automated guided vehicle 10 according to this embodiment will be described below.

[0016] FIG. 3 is a perspective view of the automated guided vehicle 10 according to this embodiment, and FIG. 4 is a plan view of the automated guided vehicle 10 according to this embodiment. As shown in FIGS. 3 and 4, the automated guided vehicle 10 according to this embodiment mainly includes a front camera 11, a rear camera 12, a measurement LiDAR 13, a connecting device 14, a main body 15 incorporating a robot control unit 20, a front LiDAR 16, a rear LiDAR 17, a bumper 18, and a pair of main wheels 19a, 19b and auxiliary wheels 19c to 19f. In this embodiment, the side having the connecting device 14 for connecting to the cart 40 is referred to as the rear, and the side having the front camera 11 is referred to as the front for explanation.

[0017] The front camera 11 is installed on the front upper part of the main body 15, captures an image of the front of the vehicle, and transmits the captured color image to the robot control unit 20 (see FIG. 6) built into the main body 15. The rear camera 12 captures an image of the rear of the vehicle, and transmits the captured color image to the robot control unit 20. The front camera 11 and the rear camera 12 are color image sensors that receive natural light or lighting light in a warehouse, which is visible light, to form an image of a subject, and for example, a CMOS image sensor or a CCD image sensor can be used. In this embodiment, a stereo camera is used as the front camera 11, and an RGB-D camera (RGB depth camera) is used as the rear camera 12. Note that the combination of the front camera 11 and the rear camera 12 is not limited to the illustrated configuration, and any combination of 3D cameras can be adopted. In addition, the installation positions of the front camera 11 and the rear camera 12 are not limited to the illustrated positions, and they can be installed at any positions. For example, the front camera 11 may be provided in the center of the bumper 18.

[0018] The measurement LiDAR 13 is a distance measuring sensor that is installed on the upper rear side of the main body 15, irradiates laser light, and receives reflected light of the irradiated laser light to measure the position of an object around the automated guided vehicle 10, the distance to the object, the shape of the object, and the like. In this embodiment, a 3D-LiDAR that irradiates laser light in multiple directions and scans 360° around the automated guided vehicle 10 in three dimensions is used as the measurement LiDAR 13. Furthermore, it is preferable that the measurement LiDAR 13 and the robot control unit 20 have a SLAM (Simultaneous Localization and Mapping) function that estimates the position of the automated guided vehicle 10 and maps the surroundings by scanning a 360° surrounding range and a certain height range. Note that the measurement LiDAR 13 can also be configured to serve as a forward LiDAR 16 and a rear LiDAR 17, which will be described later.

[0019] As shown in FIGS. 3 and 4, the connecting device 14 is a connecting device installed behind the main body 15 of the automated guided vehicle 10 and includes a hook 14a that opens downward, a contact surface 14b, and a lifting device 14c. The connecting device 14 connects to the connecting portion 44 of the cage car 40 by moving the hook 14a up and down by the lifting device 14c based on the control of the cart transport unit 205. Here, FIG. 5 is a diagram for explaining a method of connecting the connecting device 14 to the connecting portion 44 of the cage car 40. (A) shows the state before connection where the hook 14a of the connecting device 14 is in the raised position, and (B) shows the connected state where the hook 14a of the connecting device 14 is in the lowered position. In the present embodiment, the automated guided vehicle 10 identifies the connecting portion 44 of the cage car 40 as described later and moves forward to the position of the identified connecting portion 44 to bring the contact surface 14b of the connecting device 14 closer to the connecting portion 44. Here, as shown in FIG. 5(A), the lower end of the hook 14a of the connecting device 14 in the raised position is higher than the bar 43a which is the connecting portion 44. Then, after the robot control unit 20 moves the automated guided vehicle 10 forward to a position where the connecting device 14 can connect to the connecting portion 44 (that is, the position where the contact surface 14a contacts the bar 43a), as shown in FIG. 5(B), the hook 14a of the connecting device 14 is lowered by the lifting device 14c and engaged with the bar 43a to connect the connecting device 14 to the connecting portion 44. Note that a sensor for detecting contact with the connecting portion 44 may be provided on the contact surface 14b of the connecting device 14. Also, the configuration of the connecting device 14 constituting the connecting portion is not limited to the illustrated configuration, and a connecting device of any shape can be adopted according to the shape of the transport cart.

[0020] The front LiDAR 16 is installed at the lower front of the main body 15, irradiates laser light, and measures the position of an object around the automated guided vehicle 10, the distance to the object, the shape of the target object, etc. by receiving the reflected light of the irradiated laser light. It is a distance measuring sensor. In the present embodiment, the front LiDAR 16 is used to detect obstacles (including people) in front of the automated guided vehicle 10. Note that the front LiDAR 16 can also be replaced by other obstacle detection sensors (for example, ultrasonic sensors).

[0021] The rear LiDAR 17 is installed on the rear side of the main body 15 and below the measurement LiDAR 13. It irradiates laser light and measures the position of objects around the automated guided vehicle 10, the distance to the objects, the shape of the target object, etc. by receiving the reflected light of the irradiated laser light. It is a distance measuring sensor. In this embodiment, the rear LiDAR 17 is used to detect obstacles (including people) behind the automated guided vehicle 10.

[0022] The bumper 18 is provided on the front side of the main body 15, and buffer wheels 18a and 18b are provided at both ends. Even when the automated guided vehicle 10 collides with an obstacle during travel, the bumper 18 abuts on the obstacle first, preventing the main body 15, the main wheels 19a and 19b, etc. from being damaged.

[0023] A pair of main wheels 19a and 19b are provided on both side surfaces of the main body 15. The main wheels 19a and 19b are connected to a pair of wheel drive devices (not shown). By changing the rotational angular velocity of the main wheel 19a by the left drive device and the rotational speed of the main wheel 19b by the right drive device, it is possible to change the direction of the automated guided vehicle 10. The wheel drive device is configured to include general-purpose motors such as a direct current (DC) motor, an alternating current (AC) motor, a motor with an encoder, and a geared motor, and can rotate the main wheels 19a and 19b at an arbitrary rotational angular velocity. Swing-type auxiliary wheels 19c and 19d that are not connected to the drive device are provided near the center of the lower front side of the main body 15, and swing-type auxiliary wheels 19e and 19f that are not connected to the drive device are provided near the center of the lower rear side of the main body 15.

[0024] An emergency stop button 21 for emergency stop and a buzzer 22 for issuing an alarm are provided on the top surface of the main body 15. Also, a flashlight 23 is provided on the top surface of the main body 15 located below the measurement LiDAR 13.

[0025] The robot control unit 20 has a communication unit, a storage unit, and a calculation unit, and is a computer that executes the unmanned transportation process according to the present embodiment by executing the transportation program stored in the storage unit in the calculation unit. The robot control unit 20 is configured using ROS (Robot Operating System), which is a standard platform for robot development, and has a coordinate conversion function for mutually converting the coordinate systems of the cameras 11 and 12 and the coordinate systems of the LiDARs 13, 16, and 17. Further, in the present embodiment, the robot control unit 20 can perform wireless communication with the RM server 30 via the communication unit, and controls the operation of the unmanned carrier vehicle 10 based on the instructions of the RM server 30. Specifically, when the robot control unit 20 receives a conveyance instruction to convey the cage car 40 placed in a predetermined hooking area F from the RM server 30, it connects to the cage car 40 placed in the predetermined hooking area F based on this conveyance instruction, and executes an unmanned conveyance process of conveying the cage car 40 to the release area R.

[0026] As shown in FIG. 6, the robot control unit 20 includes a map management unit 201, an operation management unit 202, a route generation unit 203, and a sensor management unit 204, which are realized by executing a conveyance program in the calculation unit. The map management unit 201 has a map update function by SLAM (Simultaneous Localization AND Mapping). The environmental map (2D or 3D map) updated by the map update function is composed of a basic map for parts that do not change, such as the floor plan of the warehouse and basic passages, and a plurality of partial maps. The map management unit 201 grasps the position and shape information of objects that are constantly changing, such as cage cars and cargo boxes, based on signals from the sensor group (11 to 13, 16, 17), and real-time updates the partial map (peripheral map) around itself while traveling.

[0027] The operation management unit 202 controls the wheel drive device of the automated guided vehicle 10 based on the hooking instruction, enabling autonomous driving. Information on the rotation amounts of the main wheels 19a and 19b is stored as odometry information. The operation management unit 202 has a function of estimating its own position without communicating with the RM server 30 using the image information, odometry information, environmental map, and observation data of the sensor group (11 to 13, 16, 17). For the estimation of the own position, a known estimation method (for example, the Monte Carlo method) can be used.

[0028] FIG. 7 is a plan view of the warehouse according to the present embodiment. In the warehouse, a traveling path 62 is set while avoiding the columns 61. A plurality of wireless communication devices 63 are installed in the warehouse so that the automated guided vehicle 10 traveling on the traveling path 62 can perform wireless communication continuously. In FIG. 7, the hooking area is indicated by the symbol F, the release area is indicated by the symbol R, and the charging station is indicated by the symbol D. In order to easily identify the area to be worked on, a sign 64 composed of characters, figures, etc. serving as landmarks may be provided near each area. The automated guided vehicle 10 waits at a waiting place such as the charging station D, and when the robot control unit 20 receives a hooking instruction from the RM server 30 via the communication unit, the automated conveyance process is started. At least a part of the boundary between the hooking area F, the release area R, and the charging station D and other areas is provided with painting or color tape, enabling an operator to visually recognize each area and improving the detection accuracy of the area by the automated guided vehicle 10. The hooking area F is an area where the cart 40 loaded with goods is placed and is registered in the basic map in advance. The hooking area F is large enough to accommodate multiple carts 40, and as will be described later, the placement location of the cart 40 can also be managed by setting virtual lanes. The automated guided vehicle 10 targets the cart 40 placed in a specific lane based on a hooking instruction. In the hooking area F, an operator manually transports and parks the cart 40 loaded with goods. However, since there are variations in the position and orientation of the cart 40 parked in the hooking area F, each function of the carriage conveyance unit 205 described later is required for automatic connection.

[0029] The release area R is an area where the cart 40 transported from the hooking area F is detached from the automated guided vehicle 10 and is registered in the basic map in advance. The release area R is partitioned into a plurality of virtual lanes as will be described later, and the automated guided vehicle 10 transports the cart 40 to a specific virtual lane in the release area R. It is also possible to set a release area R in a different virtual lane for each cart 40. For example, when a shipping label such as an SCM label or a PD label is attached to the cart 40, the shipping label is imaged by the front camera 11 or the rear camera 12 to obtain shipping information, and by comparing the shipping information stored in the RM server 30 with the destination information, the lane of the release area R that is the destination can be determined. Here, the shipping label has a unique code and the code is barcoded and printed, and the current handling is that each label is circulated without being peeled off as it is. There may be a plurality of hooking areas F and release areas R respectively. By comparing the shipping information and the destination information stored in the RM server 30, it is possible to realize the conveyance process from N hooking areas F to N release areas R.

[0030] The route generation unit 203 generates an autonomous driving route using the environmental map stored in the storage device of the robot control unit 20. When a hooking instruction is issued from the RM server 30, the route generation unit 203 generates a route (global route) from the loading position to the loading / unloading position based on the basic map and the partial map, and corrects the surrounding route (local route) of the automated guided vehicle 10 as necessary when the partial map is updated while driving.

[0031] The sensor management unit 204 detects observation signals from the sensor group (11 to 13, 16, 17) and transmits them to the map management unit 201, the operation management unit 202, etc. When the sensor group (11 to 13, 16, 17) detects an obstacle (including a person), the operation management unit 202 that has received the detection information from the sensor management unit 204 performs control to avoid the obstacle or temporarily stop the travel if it is difficult to avoid.

[0032] The cart transport unit 205 has a target cart detection function 205a, a cart shape recognition function 205b, a connection position identification function 205c, a relative position adjustment function 205d, a connection unloading function 205e, and a release function 205f in order to execute the above-described automated guided vehicle transport process. Each function realized by the cart transport unit 205 will be described below.

[0033] (Target cart detection function 205a) The cart transport unit 205 determines whether there is a cage car 40 to be transported in the hooking area F. Here, FIG. 8 is a diagram for explaining the hooking method of the cage car 40 in the hooking area F by the automated guided vehicle 10. As shown in FIG. 8(A), when the automated guided vehicle 10 reaches the vicinity of the hooking area F, as shown in FIG. 8(B), the cart transport unit 205 acquires scan data of the area including the hooking area F from the measurement LiDAR 13. The cart transport unit 205 determines whether there is a cage car 40 to be transported by detecting an object in the shape of the cage car 40 based on the scan data of the hooking area F. Here, among the plurality of carts 40 in the hooking area F, the position information of the cart 40 to be transported has already been acquired from the RM server 30 in advance. When identifying the cart 40 to be transported, when there are a plurality of cage cars 40 in the hooking area F, the cage car 40 at a predetermined position (for example, the right end side, the left end side, or the position closest to the automated guided vehicle 10) may be set as the transport target, or a plurality of virtual lanes may be set, and the cart 40 located in a specific virtual lane may be set as the transport target.

[0034] When setting virtual lanes, a plurality of virtual lanes are set based on the coordinate information of the hooking area F. After identifying each lane as an empty lane without a cage car 40 and an occupied lane with a cage car 40 arranged, the target lane is identified. Here, which virtual lane to select from the plurality of virtual lanes (the method for identifying the target lane) is stored as control information in the RM server 30. The control information is transmitted to the automated guided vehicle 10 by the transport operation control unit 34 of the RM server 30, and thus the target cart detection function 205a identifies the target lane by the specific method specified by the control information. Also, for each tour course of the automated guided vehicle 10 and / or for each hooking area F, the target lane can be configured to be set as the leftmost or rightmost lane among the occupied lanes. In this case, the target cart detection function 205a includes information in the control information for identifying the rightmost or leftmost lane among the occupied lanes as the target lane for each tour course of the automated guided vehicle 10 and / or for each hooking area F. In the target cart detection function 205a, the reason for using the scan data of the measurement LiDAR 13 is that when using cameras 11 and 12, it is necessary to adjust the orientation of the automated guided vehicle 10 so that the hooking area F falls within the angle of view (for example, 120°), while there is no such restriction with the measurement LiDAR 13. In this embodiment, the target cart detection function 205a and the cart shape recognition function 205b are continuously executed when the automated guided vehicle is at the position of reference numeral 10b, but the cart shape recognition function 205b may be executed first when the automated guided vehicle is at the position of reference numeral 10a.

[0035] (Cart shape recognition function 205b) When it is determined that the cage cart 40 exists in the hooking area F, the cart transport unit 205 obtains a bounding box (rectangular area) that surrounds the cage cart 40 existing in the hooking area F by performing AI image processing on the two-dimensional image captured by the front camera 11, and then obtains a 3D bounding box (stereoscopic rectangular area) with depth information added by the function of the stereo camera. In this embodiment, the cage cart 40 can be recognized with high accuracy by performing image recognition of the cage cart 40 in the two-dimensional image using a learning model obtained by machine learning on images of a large number of cage carts. Note that, as an object detection algorithm for real-time detection of the cage cart 40, known methods such as FASTer R-CNN, YOLO, and SSD can be used.

[0036] In the cart shape recognition function, the front camera 11 is used because the measurement LiDAR 13 may misidentify an object such as a cabinet or a cardboard box having a rectangular parallelepiped or cubic shape similar to the cage cart 40 as the cage cart 40 when an object such as a cardboard box is present nearby. Although the measurement LiDAR 13 can accurately identify a surface, it is not suitable for object identification because it is point cloud data that does not include color information or the like. Therefore, a specification is adopted in which the front camera 11 acquires object identification and position information, and after moving the unmanned carrier vehicle 10 to a position substantially in front of the cage cart 40, the LiDAR 13 identifies the connection surface and connection location of the cage cart 40.

[0037] (Connection position identification function 205c) Based on the position information of the cage cart 40 obtained by the cart shape recognition function 205b, the unmanned carrier vehicle 10 moves to a position substantially in front of the cage cart 40 placed in the target lane (see reference numeral 10c in FIG. 7). Subsequently, the cart transport unit 205 acquires detailed position information (including the relative angle with the unmanned carrier vehicle 10) of the connection surface 41 of the cage cart 40 based on the scan data of the measurement LiDAR 13. Specifically, the three-dimensional coordinate information of the three-dimensional rectangular area corresponding to the target cage cart 40 obtained by the above-described cart shape recognition function 205b is converted into the three-dimensional coordinate information of the measurement LiDAR 13 by the ROS function, and the measurement LiDAR 13 is made to scan the target cage cart 40, and the distance measurement data for each part of the target cage cart 40 is acquired as point cloud data. Subsequently, the cart transport unit 205 analyzes the point cloud data detected by the measurement LiDAR 13 and identifies one side surface of the cage cart 40 as the connection surface 41.

[0038] Subsequently, a part of the identified connection surface 41 is specified as a connection point 44 for connecting the connection device 14 of the automated guided vehicle 10. In the present embodiment, the connection points 44 for each type of the cage car 40 are set as control information in the RM server 30, and the cart transport unit 205 specifies the connection point 44 on the connection surface 41 of the specified type of cage car 40 based on the control information acquired from the RM server 30. For example, the connection surface 41 of the cage car 40 shown in FIG. 2 is configured in a lattice pattern by poles 42 and bars 43. In the control information possessed by the RM server 30, together with the shape data of the cage car 40, a bar 43a extending horizontally at a height comparable to that of the connection device 14 between poles 42a and 42b near the center of the connection surface 41 is set as the connection point 44. By using the shape data of the cage car 40 included in the control information received from the RM server 30, the cart transport unit 205 can accurately specify the lowermost bar 43a located between the poles 42a and 42b as the connection point 44. Note that different connection points 44 may be set for each circuit course and / or each type of transported item, and the cart transport unit 205 may be made to specify different connection points 44 for each circuit course and / or each type of transported item based on the shipping information acquired from the shipping label.

[0039] Subsequently, the cart transport unit 205 calculates the relative position of the automated guided vehicle 10 with respect to the cage car 40 and the relative angle of the automated guided vehicle 10 with respect to the cage car 40. FIG. 10 is a diagram for explaining a method of calculating the relative position and the relative angle of the automated guided vehicle 10 with respect to the cage car 40. As shown in FIG. 10, with the connection point 44 as the origin and the orthogonal direction of the connection surface 41 (the normal line L1 shown in FIG. 10) as the X-axis, the coordinates (X, Y) of the center position of the automated guided vehicle 10 in the coordinate system are calculated as the relative position between the cage car 40 and the automated guided vehicle 10. Further, the angle θ formed by the front direction of the automated guided vehicle 10 (for example, the center line L2 of the automated guided vehicle 10 shown in FIG. 10) and the X-axis (or the normal line L1 of the connection surface 41) is calculated as the relative angle between the cage car 40 and the automated guided vehicle 10.

[0040] (Relative position adjustment function 205d) The cart transport unit 205 adjusts the relative position and relative angle (orientation) of the automated guided vehicle 10 by controlling the path along which the automated guided vehicle 10 moves toward the target cage cart 40. Here, as described above, in order to connect the hook 14a to the connection point 44, it is necessary to insert the hook 14a between the pole 42a and the pole 42b. However, in the present embodiment, the cage cart 40 uses a commercially available cage cart, and the width of the connection point 44 cannot be changed. On the other hand, the hook 14a requires a certain width in order to have the strength to transport the cage cart 40. As a result, the margin for inserting the hook 14a (the width of the connection point 44 - the width of the hook 14a) may be as small as several centimeters or less. Therefore, when the automated guided vehicle 10 approaches the cage cart 40, if the position of the connection device 14 of the automated guided vehicle 10 and the position of the connection point 44 of the cage cart 40 are misaligned in the width direction of the connection surface 41, or if the approach angle of the automated guided vehicle 10 is misaligned, there is a risk that the hook 14a cannot be inserted between the pole 42a and the pole 42b.

[0041] Therefore, when the automated guided vehicle 10 approaches the cage cart 40, the cart transport unit 205 adjusts the relative position of the automated guided vehicle 10 so that the position of the connection device 14 of the automated guided vehicle 10 and the position of the connection point 44 of the cage cart 40 coincide in the width direction of the connection surface 41, and adjusts the orientation of the automated guided vehicle 10 so that the relative angle between the automated guided vehicle 10 and the connection surface 41 of the cage cart 40 (the angle θ formed by the normal line L1 of the connection surface 41 of the cage cart 40 in FIG. 10 and the center line L2 of the automated guided vehicle 10) becomes 0 degrees. FIG. 11 is a diagram for explaining a method of adjusting the position and orientation of the automated guided vehicle 10 with respect to the carrier vehicle 40 in a scene where the automated guided vehicle 10 is connected to the carrier vehicle 40. As shown in FIG. 11, when in the position (a) facing obliquely with respect to the connection surface 41 of the carrier vehicle 40, the carriage transport unit 205 moves the automated guided vehicle 10 to a position where the positions of the connection device 14 and the connection point 44 coincide in the width direction of the connection surface 41, and adjusts the relative orientation of the automated guided vehicle 10 to an orientation in which the connection device 14 is substantially parallel to the connection surface 41, so that the connection device 14 of the automated guided vehicle 10 faces the front of the connection point 44 and is positioned at the position (b). Thereby, by simply moving the automated guided vehicle 10 straight ahead by a predetermined distance toward the carrier vehicle 40, it becomes possible to insert the hook 14a between the poles 42a and 42b of the connection surface 41 of the carrier vehicle 40.

[0042] (Connection and unloading function 205e) After the carriage transport unit 205 executes the relative position adjustment function 205d, it moves the automated guided vehicle 10 straight ahead to a position where the contact surface 14b of the connection device 14 contacts the connection point 44. When it reaches a position where the contact surface 14b substantially contacts the connection point 44, the hook 14a is lowered to connect the connection device 14 to the connection point 44 (see FIG. 8(C)). Subsequently, the carriage transport unit 205 transports the carrier vehicle 40 connected to the connection device 14 from the hooking area F toward the release area R (see FIG. 8(D)). Here, as shown in FIGS. 8(C) and 8(D), when the carrier vehicle 40 is unloaded from the hooking area F, it is moved from a predetermined one direction (the lower right direction in the example shown in FIG. 8(C)) to the position of the carrier vehicle 40. After the automated guided vehicle 10 is connected to the carrier vehicle 40, by exiting the automated guided vehicle 10 from the opposite direction (the lower left direction in the example shown in FIG. 8(D)), it is possible to prevent collision with other automated guided vehicles 10.

[0043] (Release function 205f) FIG. 9 is a diagram for explaining a method of releasing the cart 40 in the release area R by the automated guided vehicle 10. As shown in FIG. 9(A), when the automated guided vehicle 10 that pulls the cart 40 reaches a position where the measurement LiDAR 13 can scan the release area R, the cart transport unit 205 acquires scan data of the area including the release area R by the measurement LiDAR 13. Subsequently, based on the scan data of the release area R, as shown in FIG. 9(B), the cart transport unit 205 sets virtual lanes in the release area R, and identifies each lane as an occupied lane where the cart 40 is located, an empty lane where the cart 40 is not present, and an empty lane adjacent to the occupied lane as a target lane. Subsequently, as shown in FIG. 9(C), the cart transport unit 205 moves the automated guided vehicle 10 to the target lane, stops after transporting the cart 40 into the target lane, and raises the hook 14a of the coupling device 14 to release the connection with the cart 40. Subsequently, as shown in FIG. 9(D), the cart transport unit 205 causes the automated guided vehicle 10 whose connection with the cart 40 has been released to exit the release area R. Here, as shown in FIGS. 9(C) and 9(D), the automated guided vehicle 10 is moved from a predetermined direction (the lower right direction in the example shown in FIG. 9(C)) to the target lane, and after releasing the cart 40, the automated guided vehicle 10 is made to exit from the opposite direction (the upper right direction in the example shown in FIG. 9(D)), thereby preventing it from colliding with other automated guided vehicles 10.

[0044] (Automated Guided Vehicle Processing Flow) With reference to FIG. 12, the automated guided vehicle processing flow of the automated guided vehicle system 1 according to the present embodiment will be described. FIG. 12 is a flowchart showing the automated guided vehicle processing of the automated guided vehicle system 1 according to the present embodiment. In the following, it will be described on the assumption that the automated guided vehicle 10 starts the automated guided vehicle processing based on the transport instruction of the RM server 30.

[0045] In step S101, the operation management unit 202 of the robot control unit 20 functions to move the automated guided vehicle 10 to the vicinity of the hooking area F. In this embodiment, when the robot control unit 20 receives information on the hooking area F from the RM server 30, the route generation unit 203 generates a route, and the operation management unit 202 can perform travel control of the automated guided vehicle 10 so that the automated guided vehicle 10 moves to a position where the hooking area F received is within the viewing angle (FOV) of the front camera 11 and the hooking area F fits (which is also a position where the hooking area F can be scanned by the measurement LiDAR 13).

[0046] In step S102, the target carriage detection function 205a of the robot control unit 20 determines whether there is a cage car 40 in the hooking area F. Specifically, when the automated guided vehicle 10 moves to a position where the hooking area F fits within the viewing angle (FOV) of the front camera 11, the carriage transport unit 205 acquires scan data of the hooking area F by the measurement LiDAR 13, and based on the scan data, determines whether there is a cage car 40 in the cage car placement area (target lane) of the hooking area F. If there is a cage car 40 in the hooking area F (step S102 = Yes), the process proceeds to step S103. On the other hand, if there is no cage car 40 in the hooking area F (step S102 = No), the process proceeds to step S111, and a process of moving the automated guided vehicle 10 to the waiting location is performed, and the automated guided vehicle transportation process ends.

[0047] In step S103, the carriage shape recognition function 205b of the robot control unit 20 performs image recognition of the cage car 40 placed in the hooking area F. Specifically, the carriage shape recognition function 205b performs image recognition on the captured image captured by the front camera 11 using a learning model obtained by machine learning of the cage car, thereby obtaining a three-dimensional rectangular area (3D bounding box) of all the cage cars 40 placed in the hooking area F.

[0048] In step S104, based on the three-dimensional rectangular area obtained in step 103, the cage car 40 to be transported is specified. When only one cage car 40 exists in the hooking area F, the cage car 40 is set as the conveyance target. When a plurality of cage cars 40 exist in the hooking area F, the cart conveyance unit 205 sets a plurality of virtual lanes in the hooking area F, classifies each virtual lane into an occupied lane, a target lane, or an empty lane, and discriminates the cage car 40 existing in the target lane as the cage car 40 to be conveyed. Here, which lane among the occupied lanes is to be the target lane is determined based on the control information received from the RM server 30. In the present embodiment, the hooking area F is divided into a plurality of virtual lanes to specify the cage car 40 to be conveyed. However, as described above, the cage car 40 at a predetermined position (for example, the right end side, the left end side, or the position closest to the automated guided vehicle 10) may be specified as the conveyance target. When step S104 ends, the automated guided vehicle 10 moves toward the front position of the cage car 40 to be conveyed, and when it reaches a position substantially in front of the cage car 40, it executes step S105.

[0049] In step S105, the connection surface 41 and the connection point 44 of the cage car 40 to be conveyed are specified based on the scan data of the measurement LiDAR 13. Here, the acquisition of the scan data by the measurement LiDAR 13 may be continuously performed while traveling, or a three-dimensional rectangular area of the cage car 40 is specified at the front position based on the captured image of the front camera 11 at a position substantially in front of the cage car 40, and the scan data of the measurement LiDAR 13 is acquired again after converting the three-dimensional coordinate information for the measurement LiDAR 13 by the function of ROS. The connection position specifying function 205c specifies the connection point 44 after specifying the connection surface 41 of the cage car 40.

[0050] In step S106, the connection position specifying function 205c of the robot control unit 20 specifies the positional relationship including the relative angle between the carrier basket 40 to be transported and the automated guided vehicle 10 based on the scan data of the measurement LiDAR 13. Specifically, as shown in FIG. 10, the connection position specifying function 205c calculates the coordinates (X, Y) of the center position of the automated guided vehicle 10 in a coordinate system with the connection point 44 as the origin and the orthogonal direction (normal direction) of the connection surface 41 as the X-axis as the relative position between the carrier basket 40 and the automated guided vehicle 10. Further, the angle θ formed by the front direction of the automated guided vehicle 10 and the X-axis is calculated as the relative angle between the carrier basket 40 and the automated guided vehicle 10.

[0051] In step S107, based on the relative position and relative angle of the automated guided vehicle 10 with respect to the carrier basket 40 calculated in step S106, the relative position adjusting function 205d of the robot control unit 20 adjusts the position and orientation of the automated guided vehicle 10 with respect to the carrier basket 40. Specifically, as shown in FIG. 11(B), when moving the automated guided vehicle 10 to a position where the positions of the connection device 14 and the connection point 44 coincide in the width direction of the connection surface 41, the relative position adjusting function 205d adjusts the orientation of the automated guided vehicle 10 to be perpendicular to the connection surface 41 (the orientation where the angle θ is 0 degrees).

[0052] In step S108, the connection and unloading function 205e of the robot control unit 20 connects the automated guided vehicle 10 and the carrier basket 40. Specifically, the connection and unloading function 205e moves the automated guided vehicle 10 to a position where the connection device 14 can be connected to the connection point 44, and lowers the hook 14a at that position to connect the connection device 14 to the connection point 44. In step S109, the connection and unloading function 205e moves the automated guided vehicle 10 connected to the carrier basket 40 from the docking area F to the vicinity of the release area R. Subsequently, the release function 205f acquires the scan data of the release area R by the measurement LiDAR 13, sets a virtual lane in the release area R as described above, and specifies an empty lane without the information carrier basket 40 as the target lane according to the control received from the RM server 30.

[0053] In step S110, the release function 205f transfers the cage car 40 to the target lane and releases it in the release area R. Specifically, the automated guided vehicle 10 is moved to the target lane specified in step S109, stopped, and then the hook 14a is raised to disconnect it from the connection point 44, thereby releasing the cage car 40. In step S111, the automated guided vehicle 10 is moved to a preset waiting location by the function of the operation management unit 202 of the robot control unit 20.

[0054] As described above, in the automated guided vehicle system 1 according to the present embodiment, the automated guided vehicle 10 that is connected to the cage car 40 and automatically transports it to the destination has a connection device 14 for connecting to the cage car 40, a front camera 11 for imaging the cage car 40, a measurement LiDAR 13 for receiving laser light and detecting the distance to the cage car 40, and a robot control unit 20 for controlling the automated guided vehicle 10. The robot control unit 20 performs image recognition of the cage car 40 based on the captured image of the front camera 11, calculates the positional relationship between the automated guided vehicle 10 and the cage car 40 based on the detection result of the measurement LiDAR 13 corresponding to the cage car 40 recognized by the image recognition, and moves the automated guided vehicle 10 based on the calculated positional relationship, thereby connecting the connection device 14 of the automated guided vehicle 10 to the connection point 44 of the cage car 40. Here, when connecting the automated guided vehicle 10 to the cage car 40, if only the distance measurement result of the measurement LiDAR 13 is used, the approximate shape of the surrounding objects can be grasped, but there may be cases where it cannot be recognized whether the object is the cage car 40 or a box such as a cabinet or cardboard box with the same shape as the cage car 40. However, in the present embodiment, the use of the captured image of the front camera 11 can surely identify the cage car 40.

[0055] Also, when connecting the automated guided vehicle 10 to the cage car 40, if only the captured image of the front camera 11 is used, it may not be possible to accurately detect the relative position and relative angle of the automated guided vehicle 10 with respect to the cage car 40. In this case, it may be difficult to insert the hook 14a into the narrow gap between the poles 42a and 42b where the connection point 44 is located, and there may be a case where the hook 14a cannot be connected to the connection point 44. On the other hand, in the present embodiment, when recognizing the cage car 40, image recognition is performed using the captured image captured by the front camera 11. On the other hand, when calculating the positional relationship between the automated guided vehicle 10 and the cage car 40, by using the distance measurement result of the LiDAR 13 for measurement, the cage car 40 can be recognized with high accuracy, and the positional relationship between the automated guided vehicle 10 and the cage car 40 can be calculated with high accuracy. In particular, in the present embodiment, based on the detection result of the LiDAR 13 for measurement, by adjusting the relative position and relative angle of the automated guided vehicle 10 with respect to the cage car 40, even when the dimensional difference between the poles 42a and 42b and the hook 14a is small (for example, a difference of several centimeters), the hook 14a can be inserted between the poles 42a and 42b.

[0056] Further, in the present embodiment, the robot control unit 20 calculates a three-dimensional rectangular area of the recognized cage car 40 using the captured image captured by the front camera 11 which is a stereo camera, and based on the calculated three-dimensional rectangular area and the point cloud data which is the detection result of the LiDAR 13, by specifying the connection surface 41 and the connection point 44 of the cage car 40, the positional relationship between the cage car 40 and the automated guided vehicle 10 can be calculated with higher accuracy.

[0057] As described above, the preferred embodiments of the present invention have been described. However, the technical scope of the present invention is not limited to the description of the above embodiments. Various changes and improvements can be made to the above embodiments, and forms with such changes or improvements are also included in the technical scope of the present invention.

[0058] For example, the operation using the automated guided vehicle according to the present invention is not limited to the inbound and outbound operations, and can be applied to various operations that require the automated guided vehicle 10 to transport goods. In the above-described embodiment, a configuration in which the cage vehicle 40 is imaged by the front camera 11 is disclosed. However, a configuration in which the cage vehicle 40 is imaged by the rear camera 12 may also be adopted. Further, although a configuration having two cameras, i.e., the front camera 11 and the rear camera 12, is illustrated, the number of cameras is not particularly limited, and a configuration having cameras in addition to the front camera 11 and the rear camera 12 may also be used.

[0059] In addition, although a configuration having a plurality of LiDARs is disclosed, the front LiDAR and the rear LiDAR may be combined and used by an omni LiDAR that scans the entire 360° range. In the above-described embodiment, a coupling device that engages by lowering the hook 14a from above is disclosed. However, the present invention is not limited to this aspect, and a coupling device having an arbitrary engagement mechanism (for example, a coupling device having a pair of opening and closing claws) can be used.

[0060] Furthermore, in the above-described embodiment, a configuration in which an operator transports the cage vehicle 40 to the hooking area is illustrated. However, the present invention is not limited to this configuration, and the unmanned transport vehicle 10 travels to the point where the operator loads the luggage onto the cage vehicle 40, connects the unmanned transport vehicle 10 to the cage vehicle 40 at that point, and transports the cage vehicle 40 to the release area or a predetermined destination. Also, the place where the cage vehicle 40 is released is not limited to a predetermined release area. For example, an operator or a manager can individually set the transport destination of the cage vehicle 40 using the management terminal 50 or the like, and the unmanned transport vehicle 10 can be configured to transport the cage vehicle 40 to the set transport destination.

[0061] In addition, in the above-described embodiment, the unmanned transport vehicle 10 is illustrated as waiting at a waiting place such as a charging station until it receives a transport instruction from the RM server 30. However, the present invention is not limited to this configuration, and the unmanned transport vehicle 10 may be configured to travel in a predetermined route and move to a point where the cage vehicle 40 is located when it receives a transport instruction from the RM server 30.

Explanation of Reference Numerals

[0062] 1…Automated Guided Vehicle System 10…Automated Guided Vehicle 11…Front Camera 12…Rear Camera 13…LiDAR for Measurement 14…Coupling Device 15…Main Body 16…LiDAR for Front 17…LiDAR for Rear 18…Bumper 19…Wheel 20…Robot Control Unit 21…Emergency Stop Button 22…Buzzer 30…RM Server 31…Automated Guided Vehicle Management Unit 32…Operation Adjustment Unit 33…Workplace Management Unit 40…Cage Truck (Transport Cart) 41…Coupling Surface 42…Pole 43…Bar 44…Coupling Location 50…Management Terminal 61…Column 62…Passageway 61…Wireless Communication Device

Claims

1. An automated guided vehicle that automatically transports a transport cart to a destination, comprising: a coupling device for coupling to the transport cart; a camera for imaging the transport cart; a distance measuring sensor that irradiates and receives laser light to measure the distance and shape of the transport cart; a control unit, and having, The control unit includes a cart shape recognition function for recognizing the shape of a target transport cart based on the captured image of the camera, a connection position specifying function for specifying the connection location of the target transport cart based on the measurement data of the distance measuring sensor, a relative position adjustment function for moving the automated guided vehicle so that the connection location and the coupling device have a desired relative angle, a connection and unloading function for connecting the coupling device to the connection location and transporting the transport cart to the destination, and a release function for releasing the connection of the coupling device at the destination. Automated guided vehicle.

2. The control unit stores map information including a hooking area, The automated guided vehicle according to claim 1, further comprising a target cart detection function for detecting whether or not the target transport cart exists in the hooking area based on the measurement result of the distance measuring sensor.

3. When the target cart detection function of the control unit hooks a transport cart arranged in a predetermined hooking area, the hooking area is divided into a plurality of lanes, and one of the lanes in which the transport cart exists is set as a target lane in which the target transport cart exists. The automated guided vehicle according to claim 2.

4. The cart shape recognition function of the control unit has a function of performing image recognition processing on the image captured by the camera to calculate a three-dimensional rectangular area of the transport cart, The relative position adjustment function of the control unit adjusts the relative angle of the automated guided vehicle with respect to the transport cart by specifying the connection surface of the transport cart and the connection location on the connection surface based on the calculated three-dimensional rectangular area and the measurement data of the distance measuring sensor. The automated guided vehicle according to claim 1.

5. The image captured by the camera is a three-dimensional image, The measurement data of the distance measuring sensor is point cloud data. The automated guided vehicle according to claim 4.

6. The camera is a stereo camera, The distance measuring sensor is an omnidirectional LiDAR. The automated guided vehicle according to claim 5.

7. The control unit stores map information including a release area, When the release function of the control unit releases the transport cart to a predetermined release area, the release area is divided into a plurality of lanes, and one of the lanes in which the transport cart does not exist is set as the destination among the lanes. The automated guided vehicle according to claim 1.

8. The automated guided vehicle according to claim 1, wherein the connecting device includes a hook that engages with the connecting portion, a contact surface that contacts the connecting portion, and a driving device that drives the hook.

9. Furthermore, it is provided with a wireless communication device that receives control information including the vehicle type information of the transport cart and the connection location information for each vehicle type from an external server. The cart shape recognition function recognizes the shape of the target transport cart based on the control information and the captured image of the camera. The connection position specifying function specifies the connection location of the target transport cart based on the control information and the distance measurement result of the distance measurement sensor. The automated guided vehicle according to any one of claims 1 to 8.

10. A server configured to be communicable with the automated guided vehicle according to claim 9, The server that transmits the vehicle type information of the transport cart used in the cart shape recognition function and the connection location information for each vehicle type of the transport cart used in the connection position specifying function to the automated guided vehicle.

11. A management terminal configured to be communicable with the server according to claim 10, The management terminal that transmits the vehicle type information of the transport cart used in the cart shape recognition function and the connection location information for each vehicle type of the transport cart used in the connection position specifying function to the server.

Images