Transfer station with improved detection of goods held by a robot, and associated method and order-picking system

Abstract

The invention relates to a transfer station (1, 1a..1c) for transferring goods (2), the transfer station comprising a provision area (A) at a first position and a delivery area (B) at a second position. The transfer station (1, 1a..1c) also comprises a robot (3a, 3b) having a robot base (4a, 4b) and a gripper (5a, 5b). The robot (3a, 3b) or its gripper (5a, 5b) picks up the goods (2) at the first position, holds them during movement from the first position to the second position, and delivers them at the second position. The transfer station (1, 1a..1c) also comprises: a camera (6a..6c), which can be moved synchronously with the gripper (5a, 5b), for capturing an image (14, 14') of the goods (2); a control unit (7) for triggering the synchronously moving camera (6a..6c) while the goods (2) are being moved; and an object detection module (8) for detecting the goods (2) in the captured image (17). The invention also relates to an order-picking system (17) and to a method for operating such a transfer station (1, 1a..1c). The invention also relates to a robotic system for use in a transfer station (1, 1a..1c).

Description

[0001] TRANSFER STATION WITH IMPROVED DETECTION OF GOODS HELD BY A ROBOT AND ASSOCIATED METHOD AND PICKING SYSTEM

[0002] The invention relates to a transfer station for transferring goods, in particular in a picking system, which station comprises a staging area for staging goods at a first position and a delivery area for delivering the goods at a second position, which is spaced from the first position. In addition, the transfer station comprises a robot, in whose operative area the staging area and the delivery area lie and which is configured to move the goods from the first position to the second position. For this purpose, the robot has a robot base and a gripper, which is movable relative to the staging area and the delivery area and is configured to pick up the goods at the first position, hold them during a movement along a (predetermined) trajectory from the first position to the second position, and deliver them at the second position.Finally, the transfer station includes a camera for taking at least one picture of the goods.

[0003] The invention also relates to a picking system, in particular for processing picking orders, which comprises a storage area for storing goods and a transfer station of the aforementioned type. Furthermore, the picking system comprises a conveyor system for transporting the goods, which is connected to the staging area and the delivery area.

[0004] The invention also relates to a method for transferring goods using a robot of a transfer station of the aforementioned type, particularly in a picking system. The method comprises providing the goods in the staging area at the first position, picking up the goods with the gripper at the first position, moving the goods along a (predetermined) trajectory from the first position to the second position, and depositing the goods at the second position, as well as taking at least one image of the goods with the camera.

[0005] Finally, the invention relates to a robot system for use in a transfer station for transferring goods, in particular in a picking system, wherein the robot system is configured to move the goods provided in a first position in a provision area of ​​the transfer station from the first position to a second position in a delivery area of ​​the transfer station, which second position is spaced from the first position. The robot system comprises a robot base and a gripper configured to pick up the goods at the first position, hold them during a movement along a trajectory from the first position to the second position, and deliver them at the second position. In addition, the robot system comprises a camera for recording at least one image of the goods.

[0006] Such a transfer station, such a picking system, such a method, and such a robot system are generally known from the state of the art. For example, goods are transferred from a delivered carton to a loading device at an incoming goods area, or from a loading device to a shipping loading device at an outgoing goods area. To monitor the transfer process and determine whether goods have been successfully transferred, cameras are often installed in the immediate vicinity of the robot or on the robot itself.

[0007] Examples of cameras attached to a robot can be found in DE 10 2021 210 903 A1, DE 11 2017 007 025 T5, EP 3 738 719 A1 and US 11,117,262 B2.

[0008] Object recognition (goods recognition) is often performed based on an image captured by the camera, for example, to identify the goods that should be picked up from the first position. This is problematic if the robot picks up a different object instead and this error goes undetected, as this could result in goods being stored in the wrong location in the warehouse or the wrong goods being loaded onto a shipping loading device. Another problem is the unintentional picking of an undesired number of goods, for example, picking multiple items when only one item should be picked.

[0009] One object of the invention is therefore to provide an improved transfer station, an improved order-picking system, and an improved method for transferring goods. In particular, it should be possible to identify goods held by a gripper.

[0010] The object of the invention is achieved by a transfer station of the type mentioned at the outset, in which the camera can be moved synchronously with the gripper and is set up to record at least one image of the goods held by the gripper, a controller is provided which is set up to trigger the recording of the at least one image with the synchronously moved camera while the goods are moving along the trajectory, and an object recognition module is provided which is set up to recognize the goods in the at least one image.

[0011] The object of the invention is also achieved by a picking system of the type mentioned above, in which the transfer station has the type mentioned above.

[0012] The object of the invention is also achieved by a method of the type mentioned at the outset, in which the camera is moved synchronously with the gripper and at least one image of the goods is recorded while the goods are held by the gripper, in which the recording of the at least one image with the synchronously moved camera is triggered by a controller while the goods are moved along the trajectory, and in which the goods are recognized in the at least one image mentioned with the aid of an object recognition module.

[0013] Finally, the object of the invention is achieved by a robot system of the type mentioned at the outset, in which the camera can be moved synchronously with the gripper and is configured to record at least one image of the goods held by the gripper, which comprises a controller configured to trigger the recording of the at least one image with the synchronously moved camera while the goods are moving along the trajectory, and which comprises an object recognition module configured to recognize the goods in the at least one image.

[0014] The proposed measures allow goods held by a gripper to be identified while the goods are being moved by the gripper along a trajectory. This allows the goods picked up by the gripper, or the goods subsequently picked up by the gripper, to be identified, measured (length and / or width and / or height), classified, or simply counted in the subsequent process flow. If necessary, they can be identified or recognized as incorrectly picked up or defective goods. This subsequently allows error handling measures to be initiated at an early stage. This can further reduce the overall error rate due to incorrect storage of goods in the warehouse or in a shipping loading device.Advantageously, this also allows for the simple automated determination of goods dimensions (length, width, and height) during the reloading process. The picking system thus becomes independent of goods master data that would otherwise be stored in a database, or can supplement such master data in the database. It is also conceivable that the synchronously moving camera could detect when goods accidentally fall from the gripper during movement, and in particular, where they fall, in order to be able to locate them more easily for subsequent retrieval.

[0015] It is noted at this point that the term “discharging” a product in the context of this disclosure includes both “placing” and “throwing” a product at the second position.

[0016] The conveyor system can be configured to offer a first loading aid in the staging area, in which the goods are stored, and to offer a second loading aid in the delivery area, into which the goods are delivered.

[0017] Goods can be transported between the storage area and the staging area. This can include goods that are retrieved from the storage area for a picking order, transferred to the second loading device at and via the transfer station, and transported with the second loading device for further fulfillment (consolidation and / or packaging and shipping) of the picking order.

[0018] The goods can be transported between the drop-off area and a goods issue. However, the goods can also be transported from a drop-off area of ​​a first transfer station to a staging area or drop-off area of ​​a second transfer station. In other words, the goods can be transported from a drop-off area to a processing area. This can involve goods that are transported from the transfer station to the goods issue for a picking order (i.e., picking order-related). The goods can be transported to and from the transfer station with or without the use of loading equipment. Examples of loading equipment that can be considered are containers (particularly box-shaped containers, small load carriers, and the like), trays, or pallets.

[0019] The conveyor technology can include stationary conveyor technology (e.g. roller conveyors, conveyor belts and the like) and / or mobile conveyor technology, in particular conveyors with their own drive (e.g. driverless transport vehicles, autonomous mobile industrial trucks).

[0020] Furthermore, the transfer station can be a stationary transfer station. However, it is also conceivable for the transfer station to be mobile. For example, it can be mounted on a storage and retrieval machine or shuttle (self-propelled single-level or multi-level storage and retrieval vehicles) so that the goods can be transferred near or within a storage rack or a rack aisle between two storage racks. In this case, the conveyor technology can be formed by telescopic arms (which comprise controllable carriers) provided by the shuttle and movable into the rack, in particular a storage channel of the rack, or by a receiving device with a suction gripper or claw gripper. This allows a first loading aid to be provided in the supply area, in which the goods are stored, and a second loading aid to be provided in the delivery area, into which the goods are delivered (transferred).

[0021] Advantageously, the robot system is designed to be mobile, particularly rollable or self-propelled (as an automated guided vehicle (AGV) or autonomous mobile robot (AMR)). This allows the robot system to be easily moved or moved between different transfer stations as needed and used for transferring goods at these stations.

[0022] In addition, a transfer station can generally have multiple staging areas and / or drop-off areas. This allows for more complex transfer processes, for example, from a first staging area to multiple drop-off areas or from one of the multiple staging areas to one of the multiple drop-off areas.

[0023] In a conventional manner, a gripping position and / or a deposit position for the goods can be determined using the synchronously movable camera and / or one or more additional cameras, which may be arranged stationary above the staging area or above the delivery area. In particular, the additional stationary camera(s) can also be advantageously used to retrieve goods that have accidentally fallen from the gripper. The one or more additional cameras can be assigned to the transfer station or the robot system.

[0024] The controller and the object recognition module can be located in the immediate vicinity of the robot or robot system, but in principle they can also be arranged further away (spatially spaced) from it. Both the controller and the object recognition module are preferably implemented in hardware and software. It is also conceivable for the controller and / or the object recognition module to be integrated into higher-level electronic components or systems and to be part of them. The incoming / outgoing (signal or data) connections from the controller and the object recognition module can be wired or wireless. It should also be noted that the object recognition module can receive the image captured by the camera via the controller or directly from the camera.

[0025] For example, the object recognition module can include a (trained) neural network for product recognition. This network is trained during a training phase using images of the products to be recognized and can subsequently recognize known and, subsequently, unknown products in the captured image. While the use of a neural network is advantageous, other methods for product recognition can also be used in principle.

[0026] A general problem with goods detection is that the camera may also detect other objects that are not held by the gripper, but which the object recognition module can mistake for known goods. In other words, objects that are randomly detected (recorded by the camera) can be inadvertently identified as goods.

[0027] To avoid this problem, an object held by the gripper can be isolated from the background, provided the camera also captures depth information from the captured scene. However, this requires considerable technical effort, as the volume of data to be transferred to and processed by the object recognition module is comparatively large due to the spatial information. In some cases, the time typically required to transfer an item from the staging area to the delivery area may not be sufficient to identify an item in real time. Furthermore, stereo cameras and spatial depth sensors are expensive, and the captured data is not always reliable enough to isolate the object held by the gripper from the background.

[0028] It is therefore advantageous if the camera shutter speed and the speed of movement of the goods are selected such that a background captured by the camera exhibits a predefined minimum degree of blur. It is particularly advantageous if the degree of blur is a product of the camera shutter speed and the speed of movement of the goods, and if the minimum degree is greater than zero point two (0.2)° in the case of a panning movement, greater than four (4) mm in the case of a translational movement, and greater than zero point eight (0.8) mm. 0(mm x degrees) in the case of a combined movement. In this case, a combined movement can be broken down into a rotation and a translation in order to determine the aforementioned minimum dimension. The camera is preferably moved along with the gripper during the shutter speed or during a series of images (several images taken in succession in time) in such a way that the goods are always in the same position in the image, in particular approximately in the center of the image. It is also particularly advantageous if the goods always take up the same position in the image during the shutter speed or during a series of images or are always seen from the same viewing angle. If necessary, this can be achieved by appropriately controlling the robot or robot system.

[0029] In particular, the said movement of the gripper can be carried out by electrical, pneumatic, hydraulic or other, in particular mechanical, drives, for example magnetic ones, which are controlled by the controller.

[0030] The proposed measures focus object detection on the foreground of the image, which is displayed in sharp focus. Object detection in the foreground is achieved with a higher accuracy rate than in the background, provided that there are objects in the background that are known to the object detection module or at least similar to it. In other words, the blurred background makes it easier to distinguish the object to be detected from the background, and simple mathematical operations can be used to remove the background if necessary and to reliably filter out redundant image content. Especially when using (trained) neural networks, the reliability of object detection can be significantly increased by the proposed measures.A further advantage of the proposed procedure is that it is also effective when the background has only a slightly different spatial depth than the foreground and masking using the spatial depth information is not possible or produces particularly unreliable results.

[0031] In particular, the blur is selected so that objects outside the prescribed safety distance from the robot are sufficiently blurred by the camera to avoid unwanted object detection in the background. The safety distance can be, for example, two (2) meters.

[0032] To ensure a certain degree of blur, the product or gripper can also be kept no further than one (1) meter (m) from the camera during the shutter speed or during a series of images. This measure also helps ensure that the background is sufficiently blurred.

[0033] In principle, the proposed measures are already effective in a two-dimensional image of the scene captured by the camera, which minimizes the amount of data that must be transferred to and processed by the object recognition module. Therefore, goods can be identified within the time typically required for a reloading process.

[0034] It is also conceivable, however, that the camera is a stereo camera or has a depth sensor, and the object recognition module is further configured to give image areas that have a different depth than the product (i.e., an image background) less weight than image areas in which the product is depicted (i.e., an image foreground). This can further increase the hit rate for object recognition.

[0035] The transfer station can be considered a “transfer system” and named as such, especially if the controller and / or the object recognition module are located far away from the other components of the transfer station.

[0036] It should also be noted that the recording of an image or multiple images does not necessarily have to take place during the entire movement of the goods along the trajectory, but it is sufficient to record an image during the movement of the goods along part of the trajectory. Preferably, an image is recorded during a uniform movement of the goods, since movements of the goods relative to the gripper are then relatively small. However, it is also conceivable that the recording of an image takes place at the beginning of the movement, since the time for goods recognition, if this is to take place while the goods are moving along the trajectory, is comparatively long. If the goods are provided in a loading aid in the staging area, the recording of an image can preferably take place when the goods have been lifted out of the loading aid by the gripper or the loading aid has been lowered until the goods are fully visible.

[0037] Further advantageous embodiments and developments of the invention emerge from the subclaims and from the description in conjunction with the figures.

[0038] It is advantageous if the camera is mounted on the robot, especially aligned with the gripper of the robot or robot system. This ensures that the camera moves at least partially automatically in synchronization with the gripper being moved by the robot, and the goods held by the gripper are generally not obscured from the camera's field of view by parts of the robot or robot system.

[0039] It is also advantageous if the camera is mounted on a movable support at a distance from the robot, in particular between the staging area and the delivery area. The camera can be positioned in front of, behind, to the side of, above or below the robot. It is advantageous if the camera is positioned essentially centrally between the staging area and the delivery area. The camera can be moved linearly, pivoted or moved in a combined manner (moved and pivoted) in order to follow the goods or the gripper as it moves. The camera can be actively controlled by the controller and moved synchronously with the gripper. It is also conceivable for the camera to independently follow a marking on the gripper.It is also advantageous if the camera is arranged in a position elevated (at a distance from the gripper in the vertical direction) relative to the vertical position assumed by the gripper during movement along the trajectory, and if a main recording direction of the camera is inclined downwards, in particular essentially at an acute angle to a horizontal, towards the gripper, which reduces its tendency to become dirty (the camera optics). It is also particularly advantageous if the goods are rotated by the gripper during movement (during the movement process) along the trajectory in such a way that the goods are always captured by the camera from the same viewing angle, which facilitates subsequent object recognition, especially when this is based on a series of images.It is also conceivable that the gripper rotates the goods during movement along the trajectory (during at least part of the movement along the trajectory) so that the goods are captured by the camera from different angles. This increases the information available for goods recognition, which can also facilitate object recognition.

[0040] The robot can be designed, for example, as an articulated-arm robot or a gantry robot. This allows proven means for the reloading process to be used. In particular, the articulated-arm robot can have a robot arm attached to the robot base and movable relative thereto, with several arm segments movable relative to one another, with the gripper arranged at a free end of the robot arm. The controller can be configured, in particular, to pivot the robot arm during the acquisition of the at least one image, in particular about a vertical axis. A gantry robot can have several axes movable in translation relative to one another and relative to the robot base.

[0041] It is particularly advantageous if one of the articulated-arm robot's movement axes closest to the robot base is a vertical axis, and the camera is mounted on the arm segment closest to the robot base. This ensures that the camera moves at least partially automatically in synchronization with the gripper. In principle, however, the camera can also be mounted on another arm segment, particularly the one on which the gripper is located. This ensures that the camera always moves in synchronization with the gripper.

[0042] Preferably, the product is held by the gripper during image capture such that it is located approximately in the center of the image captured by the camera, or the camera coupled to the robot and movable synchronously with the gripper is preferably oriented such that the product held by the gripper is located approximately in the center of the image captured by the camera. However, it is also conceivable for the camera to be movable (in particular pivotable) relative to the arm segment to which it is attached and to be able to follow the gripper. It is also conceivable for the camera to have a zoom function to capture the product at the desired size.

[0043] In a further embodiment, one of the articulated-arm robot's movement axes closest to the robot base is again a vertical axis. The articulated-arm robot comprises a camera arm that can rotate concentrically to the vertical axis. The camera is attached to the camera arm, and the controller is configured to rotate the camera arm synchronously with the robot arm while capturing at least one image. In this case, the camera arm is driven separately, allowing the camera to be pivoted away from the gripper, for example.

[0044] It is also particularly advantageous if the robot arm segments located between a camera attachment point on the robot arm and the gripper remain fixed in position relative to each other while at least one image is being captured. This eliminates the need for the camera to actively follow the gripper, ensuring that the image of the product captured by the camera remains in the same position in the image or multiple images.

[0045] It is also advantageous if the distance between the gripper and the camera is kept constant during the shutter speed or during a series of images, as then the focus setting of the camera does not need to be changed.

[0046] In a further embodiment, the transfer station comprises a vibration detection module configured to detect vibrational movement of the goods based on a plurality of recorded images. In particular, the vibration detection module can be configured to receive goods identification data from the object detection module and to additionally detect the vibrational movement of the goods based on the goods identification data. The proposed measures make it possible to counteract excessive vibrational movement of the goods, for example when the robot's gripper is moved in antiphase to the vibrational movement of the goods. Advantageously, the position of the gripper is essentially constant in a series of images in order to facilitate object identification on the one hand and the detection of vibration on the other. It is also advantageous if at least one image is provided (linked) with a (digital) timestamp.This makes it easier to reconstruct the chronological sequence, especially in the case of image series, at a later point in time.

[0047] For a better understanding of the invention, it is explained in more detail using the following figures.

[0048] They show in a highly simplified, schematic representation:

[0049] Fig. 1 is a schematic representation of a first exemplary transfer station in

[0050] oblique view;

[0051] Fig. 2 shows the articulated arm robot of the transfer station from Fig. 1 in detailed view;

[0052] Fig. 3a shows an exemplary image taken by the camera and schematically shown of a product that is moved with the gripper;

[0053] Fig. 3b shows an exemplary image taken by the camera in a real situation of a product being moved with the gripper;

[0054] Fig. 4 a transfer station with a camera arranged at a distance from the robot in

[0055] side view;

[0056] Fig. 5 an exemplary gantry robot and

[0057] Fig. 6 is a schematic view of an exemplary picking system.

[0058] By way of introduction, it should be noted that in the variously described embodiments, identical parts are provided with identical reference symbols or component designations. The disclosure contained in the entire description can be applied analogously to identical parts with identical reference symbols or component designations. Furthermore, the position information chosen in the description, such as top, bottom, side, etc., refers to the directly described and illustrated figure and, in the event of a change in position, is to be applied analogously to the new position.

[0059] Fig. 1 shows an exemplary transfer station 1a for transferring goods 2 in an oblique view. The transfer station 1a comprises a provision area A for providing a product 2 at a first position and a delivery area B for delivering the product 2 at a second position, which is spaced from the first position. The transfer station 1a further comprises a robot 3a, in whose area of ​​action the provision area A and the delivery area B lie and which is configured to move the product 2 from the first position to the second position. The robot 3a has a robot base 4a and a gripper 5a, which is movable relative to the provision area A and the delivery area B and is configured to pick up the product 2 at the first position, hold it during a movement along a (predetermined) trajectory from the first position to the second position, and deliver it at the second position.In addition, the transfer station 1a comprises a camera 6a that can be moved synchronously with the gripper 5a and is configured to capture at least one image of the goods 2 held by the gripper 5a. Furthermore, the transfer station 1a comprises a controller 7 configured to trigger the capture of at least one image with the synchronously moved camera 6a while the goods 2 are moving along the trajectory. Finally, the transfer station 1a comprises an object recognition module 8 configured to recognize the goods 2 in the at least one image.

[0060] A method for reloading the goods 2 with the robot 3a of the reloading station 1a can now comprise the following process steps:

[0061] Providing the goods 2 in the provision area A at a first position, picking up the goods 2 with the gripper 5a at the first position, moving the goods 2 along a (predetermined) trajectory from the first position to the second position and releasing (placing or dropping) the goods 2 at the second position, synchronously moving the camera 6a with the gripper 5a and recording at least one image of the goods 2 held by the gripper 5a, wherein the recording of the at least one image with the synchronously moving camera 6a is triggered by the controller 7 during the movement of the goods 2 along the trajectory and wherein the goods 2 are recognized in said at least one image with the aid of the object recognition module 8.

[0062] For example, the object recognition module 8 can have a neural network for recognizing the goods 2. This neural network is trained during a training phase with images of goods 2 to be recognized and can subsequently recognize known goods 2 and subsequently also unknown goods 2 in the recorded image. While the use of a neural network is advantageous, other methods for recognizing goods 2 can also be used in principle.

[0063] In a manner known per se, a gripping position and / or a deposit position for the goods 2 can be determined using the synchronously moving camera 6a and / or one or more additional cameras (not shown in the figures), which may be stationary above the staging area A or above the delivery area B. It is also conceivable for the synchronously moving camera 6a to detect when the goods 2 unintentionally fall from the gripper 5a during the movement and, in particular, where they fall, in order to be able to locate them more easily for subsequent retrieval. In this case, too, a gripping position can be determined using the synchronously moving camera 6a and / or the additional camera(s).

[0064] In the example shown in Fig. 1, the goods 2 are transported to the transfer station 1a using a first loading device 9a and are provided there in the staging area A. Furthermore, the goods 2 are transported from the transfer station 1a using a second loading device 9b. While the use of loading devices 9a, 9b is advantageous, in principle the goods 2 can also be transported to and from the transfer station 1a without loading devices 9a, 9b. Furthermore, trays or pallets can also be used instead of the container-shaped loading devices 9a, 9b.

[0065] Furthermore, in the example shown in Fig. 1, the staging area A, the delivery area B, and the robot 3a are located on a frame 10. Furthermore, the transfer station 1a comprises a stationary conveyor system 11a, 11b, which is connected to the staging area A and the delivery area B. These features are also not mandatory; the transfer station 1a could also be constructed differently. In particular, the conveyor system 11a, 11b could also be formed by autonomous mobile industrial trucks (driverless transport vehicles with their own transport drive).

[0066] It should also be noted that the controller 7 and the object recognition module 8, as shown in Fig. 1, can be located in the immediate vicinity of the frame 10 or robot 3a, but could in principle also be arranged further away from them or built into the frame 10. Both the controller 7 and the object recognition module 8 are preferably constructed in hardware and software. It is also conceivable that the controller 7 and / or the object recognition module 8 are integrated into higher-level electronic components or systems and are part of the same. In Fig. 1, the connections outgoing from the controller 7 and the object recognition module 8 are shown as wired, but these can also be wireless. It should also be noted that the object recognition module 8 processes the image recorded by the camera 6a, as shown in Fig.1, can be received via the controller 7, but the object recognition module 8 can also receive the image directly from the camera 6a.

[0067] Preferably, the camera 6a is moved along with the gripper 5a during the shutter speed in such a way that the product 2 is always in the same position in the image, in particular approximately in the center of the image. In particular, it is also advantageous if the product 2 is always recorded in a series of images in essentially the same position and always from essentially the same viewing angle. This means that it is particularly advantageous if the product 2 always assumes the same position in the image during the shutter speed or during a series of images. If necessary, this can be achieved by appropriately controlling the robot 3a. It is also conceivable for the camera 6a to have a zoom function in order to be able to depict the product 2 in the desired size.

[0068] In the example shown in Fig. 1, the robot 3a is designed as an articulated-arm robot, which is now shown in detail in Fig. 2. The articulated-arm robot 3a comprises a robot arm 12, which is attached to the robot base 4a and movable relative thereto, and has a plurality of arm segments 13a..13e that can be moved relative to one another. The gripper 5a is arranged at a free end of the robot arm 12, specifically at the arm segment 13e.

[0069] It is advantageous if a movement axis C of the articulated-arm robot 3a closest to the robot base 4a is a vertical axis and the camera 6a is mounted on the arm segment 13a closest to the robot base 4a, as is the case in Figs. 1 and 2. As a result, the camera 6a is moved at least partially automatically in synchronization with the gripper 5a.

[0070] Preferably, the product 2 is held by the gripper 5a during image capture so that it is located approximately in the center of the image. The camera 6a can then be rigidly attached to the arm segment 13a. However, it is also conceivable for the camera 6a to be pivotable relative to the arm segment 13a to which it is attached and to be able to follow the gripper 5. In particular, the aforementioned pivoting movement can be carried out by electrical, pneumatic, hydraulic, or other, particularly mechanical, drives controlled by the controller 7. It is also conceivable for the camera 6a to have a zoom function in order to be able to image the product 2 in the desired size.

[0071] It is also advantageous if the arm segments 13b..13e of the robot arm 12, which are located between a mounting point of the camera 6a on the robot arm 12 and the gripper 5a, remain fixed in position relative to each other during the recording of at least one image. This eliminates the need for the camera 6a to actively follow the gripper 5a, but rather the product 2 remains in the same position in the image or in the multiple images.

[0072] In the example shown in Figs. 1 and 2, the camera 6a is attached to the arm segment 13a, but it is also conceivable that it is attached to one of the other arm segments 13b..13e, in particular to the arm segment 13e on which the gripper 5a is also arranged. Furthermore, it is conceivable that the articulated-arm robot 3a comprises a camera arm that can be rotated concentrically to the vertical axis C, the camera 6a is attached to the camera arm, and the controller 7 is configured to rotate the camera arm synchronously with the robot arm 12 while capturing the at least one image 14. In this case, the camera arm is therefore driven separately, whereby the camera 6a can also be pivoted away from the gripper 5a, for example.

[0073] Fig. 3A now shows an exemplary image 14 taken by the camera 6a and schematically displayed. In the center of the image is the product 2 held by the gripper 5a. The product 2 is located in the foreground D of the image 14 and is sharply depicted.

[0074] Preferably, a shutter speed of the camera 6a and a speed when moving the goods 2 are selected such that a background E recorded by the camera 6a has a predeterminable minimum degree of blur. For this purpose, a blurred structure 15 is shown as an example in Fig. 3A. In particular, the degree of blur can be a product of the shutter speed of the camera 6a and a speed when moving the goods 2, wherein the minimum degree is greater than 0.2° in the case of a panning movement, greater than 4 mm in the case of a translational movement, and greater than 0.8 mm. 0in the case of a combined movement. The movement mentioned refers to the movement of the camera 6a. In the case of Fig. 1 and 2, the decisive movement is therefore a pivoting movement. This means that in the example shown, the controller 7 is set up to pivot the robot arm 12 while recording the image 14. Advantageously, the proposed measures concentrate object detection on the foreground D, or object detection in the foreground D occurs with a higher hit rate than in the background E. If there are objects in the background E that are known to the object detection module 8 or are at least similar to it, they are therefore less likely to be detected.

[0075] In particular, the blur is selected such that objects outside a prescribed safety distance from the robot 3a are imaged sufficiently blurred by the camera 6a to avoid unwanted object detection in the background E. The safety distance can be, for example, 2 m.

[0076] To ensure a certain degree of blur, it can also be provided that the product 2 or the gripper 5a is not further away from the camera 6a than Im during the shutter speed or during a series of images. This measure also contributes to ensuring that the background E is depicted with sufficient blur.

[0077] It is also advantageous if the distance between the gripper 5a and the camera 6a is kept constant during the shutter speed or during a series of images, since the focus setting of the camera 6a then does not need to be changed.

[0078] In principle, the proposed measures are already effective in a two-dimensional image of the scene captured by camera 6a, thereby minimizing the amount of data that must be transferred to and processed by the object recognition module 8. However, it is also conceivable that the camera 6a is a stereo camera or has a spatial depth sensor, and the object recognition module 8 is further configured to rate image areas that have a different spatial depth than the product 2 (i.e., the background E in Fig. 3A) lower than image areas in which the product 2 is depicted (i.e., the foreground D in Fig. 3A). This can further increase the hit rate for object recognition.

[0079] In a further embodiment, the transfer station 1a comprises a vibration detection module configured to detect a vibrational movement of the goods 2 based on a plurality of recorded images 14. In particular, the vibration detection module can be configured to receive goods identification data from the object detection module 8 and to additionally detect the vibrational movement of the goods 2 based on the goods identification data. The proposed measures make it possible to counteract excessive vibrational movement of the goods 2, for example, when the gripper 5a of the robot 3a is moved in antiphase to the vibrational movement of the goods 2. Advantageously, the position of the gripper 5a is essentially constant in a series of images 14 in order to facilitate object identification on the one hand and the detection of vibration on the other.

[0080] Fig. 3B shows an additional exemplary image 14 taken by the camera 6a in a real situation. 1 of a product 2 held and moved by the gripper 5a. The statements made regarding Fig. 3A apply accordingly. In particular, the blurred structure 15 is clearly shown in Fig. 3B. 1 or the motion blur of the background E is visible.

[0081] Fig. 4 now shows an embodiment of a transfer station 1b, which is similar to the transfer station 1a shown in Fig. 1. In contrast, the camera 6b is arranged at a distance from the robot 3a between the preparation area A and the delivery area B and is movably mounted. In the present case, it is arranged opposite the robot 3a, which is advantageous but not mandatory. It is also conceivable for it to be arranged to the side of, above, below or behind the robot 3a. It is advantageous if the camera 6b is arranged substantially centrally between the preparation area A and the delivery area B. The camera 6b can be moved linearly, pivoted or moved in combination in order to follow the goods 2 or the gripper 5a during its movement.The camera 6b can be actively controlled by the controller 7 and moved synchronously with the gripper 5a, but it is also conceivable for the camera 6b to independently follow a marking arranged on the gripper 5a. Fig. 4 also shows that the camera 6b is tilted downwards, which reduces its tendency to become dirty. It is particularly advantageous if the product 2 is rotated by the gripper 5a during the movement along the trajectory in such a way that the product 2 is always captured by the camera 6b from the same viewing angle, which facilitates subsequent object recognition and / or vibration detection, particularly if this is based on a series of images. However, it is also conceivable for the product 2a to be rotated by the gripper 5a during the movement along the trajectory (during the movement along at least a section of the trajectory) in such a way that the product 2a is captured by the camera 6b from different viewing angles.This increases the information available for goods recognition, which can also facilitate object recognition. Fig. 5 now shows a further embodiment of a transfer station 1c, which is similar to the transfer station 1a shown in Fig. 1. In contrast, the robot here is designed as a gantry robot 3b with several axes that can be moved in a translational manner relative to one another and relative to the robot base 4b. The camera 6c in this case is mounted directly on the movement segment to which the gripper 5b is also attached. The camera 6c is therefore automatically moved along with the goods or the gripper 5b. In principle, however, the camera 6c can also be arranged at a different location on the gantry robot 3b, for example further away from the gripper 5b. The embodiments described above and the resulting advantages also apply analogously to the transfer station 1c with the gantry robot 3b.A resulting difference is that the camera 6c is moved translationally and is not panned at the transfer station 1a.

[0082] At this point, it should be noted that a transfer station la..lc can generally have several staging areas A and / or delivery areas B. In this way, even more complex transfer processes can be carried out.

[0083] In the examples presented so far, the transfer station la..lc is a stationary transfer station. However, it is also conceivable that the transfer station la..lc is mobile. For example, it can be mounted on a storage and retrieval machine or shuttle, allowing goods 2 to be transferred near a storage rack, for example, in an aisle between two storage racks.

[0084] Fig. 6 now shows a schematically illustrated, exemplary picking system 16, in particular for processing picking orders, which comprises a storage area 17 for storing goods 2 and a transfer station 1 for transferring goods 2, wherein the transfer station 1 can be designed, for example, as already described. In addition, the picking system comprises a conveyor system 11a, 11b for transporting the goods 2, which is connected to the staging area A and the delivery area B. In particular, the conveyor system 11a, 11b can be configured to present a first loading aid 9a together with the goods 2 contained therein in the staging area A, and to present a second loading aid 9b in the delivery area B, into which the goods 2 are delivered.

[0085] Goods 2 can therefore be transported between storage area 17 and staging area A. This can be goods 2 that are retrieved from storage area 17 for a picking order. Goods 2 can also be transported between delivery area B and a goods issue. However, goods 2 can also be transported from a delivery area B of a first transfer station 1 to staging area A or delivery area B of a second transfer station. In other words, goods 2 can be transported from a delivery area B to a processing area. This can be goods 2 that are transported from transfer station 1 to the goods issue for a picking order.

[0086] Finally, it should be noted that the scope of protection is determined by the patent claims. However, the description and drawings must be used to interpret the claims. Individual features or combinations of features from the various embodiments shown and described may represent independent inventive solutions. The problem underlying these independent inventive solutions can be derived from the description.

[0087] In particular, it is also noted that the devices depicted may in reality comprise more or fewer components than shown. In some cases, the depicted devices or their components may also be shown not to scale and / or enlarged and / or reduced in size.

[0088] Reference symbol list

[0089] 1, la..lc transfer station

[0090] 2 goods

[0091] 3 a, 3b robots

[0092] 4a, 4b Robot base

[0093] 5 a, 5b gripper

[0094] 6a..6c Camera

[0095] 7 Control

[0096] 8 Object recognition module

[0097] 9a, 9b Loading aids

[0098] 10 frame

[0099] 11a, 11b Conveyor technology

[0100] 12 Robot arm

[0101] 13a..13e arm segment

[0102] 14, 14' Image

[0103] 15, 15 1 blurred structure

[0104] 16 picking system

[0105] 17 Storage area

[0106] A staging area

[0107] B Delivery area

[0108] C Movement axis

[0109] D Foreground

[0110] E Background

Claims

P a t e n t a n s p r ü c h e 1. Transfer station (1, 1a..1c) for transferring goods (2), in particular in a picking system (16), comprising a provision area (A) for providing a good (2) at a first position, a delivery area (B) for delivering the good (2) at a second position which is spaced from the first position, a robot (3a, 3b), in whose area of ​​action the provision area (A) and the delivery area (B) are located and which is configured to move the good (2) from the first position to the second position, wherein the robot (3a, 3b) has a robot base (4a, 4b) and a gripper (5a, 5b), wherein the gripper (5a, 5b) is movable relative to the provision area (A) and the delivery area (B) and is configured to pick up the good (2) at the first position, to hold it during a movement along a trajectory from the first position to the second position and to deliver it at the second position and a Camera (6a..6c) for recording at least one image (14, 14') of the goods (2), characterized in that the camera (6a..6c) can be moved synchronously with the gripper (5a, 5b) and is set up to record at least one image (14, 14') of the goods (2) held by the gripper (5a, 5b), the transfer station (1, 1a..lc) comprises a controller (7) which is set up to trigger the recording of the at least one image (14, 14') with the synchronously moved camera (6a..6c) during the movement of the goods (2) along the trajectory, and the transfer station (1, 1a..lc) comprises an object recognition module (8) which is set up to recognize the goods (2) in the at least one image (14, 14').

2. Transfer station (1, 1a..lc) according to claim 1, characterized in that a shutter speed of the camera (6a..6c) and a speed when moving the goods (2) are selected such that a background (E) recorded by the camera (6a..6c) has a predeterminable minimum degree of blur.

3. Transfer station (1, 1a..lc) according to claim 2, characterized in that the degree of blur is a product of the shutter speed of the camera (6a..6c) and a speed when moving the goods (2) and the minimum degree is greater than 0.2° in the case of a pivoting movement, greater than 4 mm in the case of a translational movement and greater than 0.8 mm- 0 in the case of a combined movement.

4. Transfer station (1, 1a..lc) according to one of claims 1 to 3, characterized in that the camera (6a..6c) is mounted on the robot (3a, 3b).

5. Transfer station (1, 1a..lc) according to one of claims 1 to 3, characterized in that the camera (6a..6c) is arranged at a distance from the robot (3a, 3b) between the provision area (A) and the delivery area (B) and is movably mounted.

6. Transfer station (1, 1a..lc) according to one of claims 1 to 5, characterized in that the robot (3a, 3b) is designed as an articulated arm robot or as a gantry robot.

7. Transfer station (1, 1a..lc) according to claim 6, characterized in that the articulated arm robot (3a) has a robot arm (12) fastened to the robot base (4a) and movable relative thereto, with a plurality of arm segments (13a..13e) movable relative to one another, wherein the gripper (5a) is arranged at a free end of the robot arm (12).

8. Transfer station (1, 1a..lc) according to claim 7, characterized in that a movement axis (C) of the articulated arm robot (3a) closest to the robot base (4a) is a vertical axis and the camera (6a) is attached to the arm segment (12a) closest to the robot base (4a).

9. Transfer station (1, 1a..lc) according to claim 7, characterized in that a movement axis (C) of the articulated arm robot (3a) closest to the robot base (4a) is a vertical axis (C), the articulated arm robot (3a) comprises a camera arm rotatable concentrically to said vertical axis (C), the camera (6a) is attached to the camera arm and the controller (7) is configured to rotate the camera arm synchronously with the robot arm (12) during the recording of the at least one image (14, 14').

10. Transfer station (1, 1a..lc) according to one of claims 7 to 9, characterized in that arm segments (13a..13e) of the robot arm (12), which lie between a fastening point of the camera (6a) on the robot arm (12) and the gripper (5a), remain fixed in position relative to one another during the recording of the at least one image (14, 14').

11. Transfer station (1, 1a..lc) according to one of claims 7 to 10, characterized in that the controller (7) is configured to pivot the robot arm (12) during the recording of the at least one image (14, 14').

12. Transfer station (1, 1a..lc) according to claim 6, characterized in that the gantry robot (3b) has several axes which can be moved translationally relative to one another and relative to the robot base (4b).

13. Transfer station (1, la..lc) according to one of claims 1 to 12, characterized in that the object recognition module (8) has a neural network for recognizing the goods (2).

14. Transfer station (1, 1a..lc) according to one of claims 1 to 13, characterized by a vibration detection module which is designed to detect a vibration movement of the goods (2) on the basis of several recorded images (14, 14').

15. Transfer station (1, 1a..lc) according to claim 14, characterized in that the vibration detection module is configured to receive goods identification data from the object detection module (8) and to additionally detect the vibration movement of the goods (2) on the basis of the goods identification data.

16. Transfer station (1, 1a..lc) according to one of claims 1 to 15, characterized in that the camera (6a..6c) is a stereo camera or has a spatial depth sensor and the object recognition module (8) is further configured to rate image areas which have a different spatial depth than the goods (2) less than image areas in which the goods (2) are depicted.

17. Transfer station (1, la..lc) according to one of claims 1 to 16, characterized in that the transfer station (1, la..lc) is designed to be mobile, in particular as a storage and retrieval device or shuttle, with which the goods (2) can be transferred near or within a storage rack or a rack aisle between two storage racks.

18. Order picking system (16), in particular for processing order picking orders, comprising a storage area (17) for storing goods (2), a transfer station (1, 1a..lc) for transferring goods (2) and a conveyor system (11a, 11b) for transporting the goods (2), characterized in that the transfer station (1, 1a..lc) is designed according to one of claims 1 to 17 and the conveyor system (11a, 11b) is connected to the provision area (A) and to the delivery area (B).

19. Order-picking system (16) according to claim 18, characterized in that the conveyor system (11a, 11b) is designed to offer a first loading aid (9a) in the provision area (A), wherein the goods (2) are stored in the first loading aid (9a), and to offer a second loading aid (9b) in the delivery area (B), wherein the goods (2) are delivered into the second loading aid (9b).

20. Method for reloading goods (2) with a robot (3a, 3b) of a reloading station (1, 1a..1c), in particular in a picking system (16), wherein the reloading station (1, 1a..1c) has a provision area (A) for providing the goods (2) at a first position and a delivery area (B) for delivering the goods (2) at a second position, which is spaced from the first position, wherein the robot (3a, 3b) has an effective area which includes the provision area (A) and the delivery area (B) and which has a robot base (4a, 4b) and a gripper (5a, 5b), and wherein the reloading station (1, 1a..1c) has a camera (6a..6c) which can be moved synchronously with the gripper (5a, 5b), comprising the steps Providing the goods (2) in the provision area (A) at the first position, picking up the goods (2) with the gripper (5a, 5b) at the first position, moving the goods (2) along a trajectory from the first position to the second position and delivering the goods (2) in the second position, Recording at least one image (14, 14') of the goods (2) with the camera (6a..6c), characterized in that the camera (6a..6c) is moved synchronously with the gripper (5a, 5b) and the at least one image (14, 14') of the goods (2) is recorded while the goods (2) are held by the gripper (5a, 5b), the recording of the at least one image (14, 14') with the synchronously moved camera (6a..6c) is triggered by a controller (7) during the movement of the goods (2) along the trajectory and the goods (2) are recognized in said at least one image (14, 14') with the aid of an object recognition module (8).

21. Robot system for use in a transfer station (1, 1a..lc) for transferring goods (2), in particular in a picking system (16), wherein the robot system is configured to move the goods (2) provided in a first position in a provision area (A) of the transfer station (1, 1a..lc) from the first position to a second position in a delivery area (B) of the transfer station (1, 1a..lc), which is spaced from the first position, comprises a robot base (4a, 4b) and a gripper (5a, 5b) which is configured to pick up the goods (2) at the first position, hold them during a movement along a trajectory from the first position to the second position and deliver them at the second position, a camera (6a..6c) for recording at least one image (14, 14') of the goods (2), which can be moved synchronously with the gripper (5a, 5b) and is set up to record at least one image (14, 14') of the goods (2) held by the gripper (5a, 5b), comprises a controller (7) which is set up to trigger the recording of the at least one image (14, 14') with the synchronously moved camera (6a..6c) while the goods (2) are moving along the trajectory, and comprises an object recognition module (8) which is set up to recognize the goods (2) in the at least one image (14, 14').

Images