Method for organising goods in a logistics area

Abstract

The invention relates to a method for organising goods (6) in a logistics area (1) comprising at least one driverless transport device (3), wherein at least one verification area (2) is defined, and the method comprises an assignment method which has the following steps: a. detecting at least one good (6) located in the verification area (2) by means of at least one sensor arrangement of the transport device (3); b. determining the pose of the good; c. checking whether the good (6) is located in a defined location (5, 5a, 5b), wherein the check involves comparing at least the pose of the good with pose data from defined locations, wherein the defined locations, along with their pose data and occupancy data, are stored in a database; d. designating at least one defined location (5, 5a, 5b) if the check has determined that the good (6) is not located at at least one defined location (5, 5a, 5b), wherein the designated defined location (5, 5a, 5b) comprises the pose of the good, and assigning the good (6) to the designated at least one defined location (5, 5a, 5b) and storing the designated location (5, 5a, 5b) in the database.

Description

[0001]31232WO Method for organizing goods in a logistics area The invention relates to a method for organizing goods in a logistics area with at least one driverless transport device. The described method serves to utilize automatically recognized points of interest (POIs) for the interaction of a mobile autonomous robot (AGV / AMR) with its environment (e.g., picking up load carriers, approaching transfer positions, mapping objects, etc.). In the following, these POIs are summarized under the term "defined location" (= station). Such defined locations can be, for example, a storage location or storage area, an arrival location or area for new goods, a pickup location or area for goods to be transported away, a loading station for the transport device, a waiting location for transport devices without an order, etc.A defined location can be the size of a good or have additional safety or tolerance zones around it. The prerequisite for the process is the sensory detection of specific goods and the determination of their pose. The detection itself can be implemented in any way. The process is characterized by the fact that it enables mobile, autonomous robots to carry out a wide variety of tasks ad hoc. This distinguishes it from previous, known processes, which require a high degree of a priori knowledge, such as manually defined transfer positions (= locations). One of the main fields of application of the process presented is the automatic detection and positioning of goods in goods logistics / intralogistics. Current methods are limited in that individual predefined locations (stations) must be specified and predefined manually.Detection is typically only possible within a range of less than + / -20 cm on the horizontal plane (x, y coordinates) and a few degrees (<10°) of rotation around the vertical axis. For occupancy detection, a check is usually carried out to determine whether sensor data (e.g. LIDAR point cloud) matching the item is available in the permissible area around the actual location. To determine the pose, the edges / contours of the load carrier are detected and the rotation and position are determined. Current methods therefore often separate the two tasks of occupancy detection of the station and determining the pose (position and orientation in space, i.e. position + orientation) of the load carrier. The first step is to determine whether a location is occupied by an object / item that can be assigned to a certain type. The second step is to determine the position of the object / item in relation to the sensor system or the robot.This makes the previous system rigid, complex to process and maintain, and prone to errors. The object of the invention is therefore to provide a method that offers greater flexibility and greater efficiency in the recognition / detection and handling / processing of goods. This object is achieved according to the invention in that at least one inspection area is defined and the method comprises an allocation method having the following steps: a. Detection of at least one item of goods located in the inspection area by at least one sensor arrangement of the transport device; b. Determination of the pose of the item, preferably by the transport device; c.Checking whether the item is located at a defined location, wherein the check comprises comparing at least the pose of the item with pose data from defined locations, wherein the defined locations, including their pose data and occupancy data, are stored in a database; d. Creating at least one defined location if the check determines that the item is not located at at least one defined location, wherein the created defined location includes the pose of the item, assigning the item to the created at least one defined location, and saving the created location in the database. This eliminates the need to specify locations manually. The transport device recognizes the pose of the item and creates an associated location in the database, instead of being limited to a narrow uncertainty range around a predefined location, as was previously the case.The space can then be deleted again after the goods have been removed by a transport device or a higher-level unit or controller, or even a user. The specific innovation therefore consists in the dynamic creation and deletion of spaces in an area that can be many times larger than the goods themselves. The method according to the invention is usually carried out by a logistics system that comprises at least one driverless transport device. The logistics system preferably comprises at least one higher-level control unit, particularly preferably a central control unit, wherein the control unit is designed to communicate with the transport devices and to exchange data and, if necessary, to control them. The logistics system can comprise several transport devices that communicate with each other. In such a case, it can be provided that no higher-level control unit is present.The driverless transport device is usually a driverless transport vehicle (Automated Guided Vehicle "AGV", Autonomous Guided Vehicle "AGV", or Autonomous Mobile Robot "AMR") that automatically moves and relocates goods in the logistics sector. The logistics sector is a geographical area, for example a room, floor, building, area, or part of the aforementioned areas, which serves goods logistics. Examples of such logistics areas are warehouses for goods, logistics centers such as train stations, ports, especially industrial ports, halls, transshipment points, and the like. Goods are movable, material objects of value that are part of the logistics sector. Goods preferably include load carriers such as pallets or the like. This enables better handling with the transport equipment.Defining the inspection area can be a manual step performed by a user, for example by defining a specific area of ​​the logistics area as the inspection area. This can be done, for example, via an interface / HMI (Human Machine Interface), preferably by marking it on a digital map. The definition can also be an automatic step, for example by including natural boundaries (e.g. walls of a room / area or similar) and / or the boundaries of the logistics area. For example, it can be provided that the entire logistics area is defined as the inspection area by default and / or automatically. The occupancy data is data that provides information about whether the defined space is occupied and / or what the defined space is occupied with. Occupancy means that at least one item of goods is located in the space.The occupancy data can include at least one parameter of the asset and / or provide information about the type of asset. They can also include data that enable unique identification of the asset, for example a serial number of the asset. The pose of an asset or pose data is information regarding the position, size or extent (in at least one axis, preferably in at least two or particularly preferably in all three axes), shape, rotation (around at least one axis, preferably around at least two or particularly preferably around all three axes) and / or orientation (in at least one axis, preferably in at least two or particularly preferably in all three axes) of an asset or a defined location.Preferably, the determination of the pose of the item comprises determining the position of the item in the inspection area and / or logistics area and / or determining the location of the item in relation to the inspection area and / or logistics area. The pose of an item (or pose data) preferably comprises the position in the logistics area as well as the location in the logistics area. Location here refers to a spatial orientation and / or rotation. The position in space preferably comprises area or volume information of the item or the location and their arrangement in relation to the logistics area and / or inspection area. Sensor data from the sensor arrangement can be used to determine the pose of the item, preferably after they have been filtered by at least one filter. The detections can be filtered, for example, based on the inspection areas, confidence, height above the ground and / or the location in space.The determination is preferably carried out at least partially by the transport device. A central computing center / processing unit can also carry out the determination at least partially. The database can be located wholly or partially in the transport device and / or wholly or partially in a central computing unit that communicates with the transport device. It can be provided that the check to determine whether the item is located at at least one defined location includes whether the item is located completely or at least partially in the at least one defined location, and preferably the check is then evaluated as positive. In other words, it can be sufficient if the pose data of the item and the location overlap at least partially.It can also be provided that the check to determine whether the asset is located at at least one defined location includes whether at least a predefined proportion of the asset (for example 50%, 60%, or 70% of the floor space of the asset and / or the defined location) is located at at least one defined location, and preferably the check is then assessed as positive. What is said in this paragraph regarding the check can also apply to the further check. It can be provided that an asset is located at more than one location. This can particularly be the case if the asset is larger than at least one of the locations. In this case, the asset can be or will be assigned to several locations. Assigning the asset means that it is entered in the database that the asset is currently located at the defined location.For this purpose, a separate data set relating to goods can be available, whereby the entries in the data set for the defined locations can be linked to the entries in the data set relating to the goods. It can also be provided that, depending on at least one further factor, at least one defined location is not created if the check could not assign the goods to at least one defined location. In such a case, it can be provided that the transport device picks up the goods and moves them into a different position, preferably transports them to a defined location. The factor can, for example, comprise a parameter which depends on the type of goods or the loading status of the transport device. For example, it can be provided that an empty pallet is taken directly to a pallet storage location when it is detected. The creation of the defined location is then not necessary.Preferably, if the check could not assign the item to at least one defined location, at least one picking pose for an item is determined, which is assumed / approached by a transport device during or before picking up the item in question, and this picking pose is preferably stored in the database. This picking pose can, for example, comprise a point, a direction, or an area that is assumed or traveled along by a transport device during or before picking up the item in question. In the case of a pallet, for example, the picking pose comprises the point at which the transport device, with suitable alignment, can begin to insert the fork / lifting fork into the provided openings in the pallet. More than one picking position can also be provided, for example in the case of symmetrical pallets.It is preferably provided that the goods are detected while the transport device is moving. The transport device can perform this task incidentally during a journey, which makes the detection of new goods more efficient. It is preferably provided that at least one transport device carries out at least one search journey during which it moves in the inspection area or along the inspection area, and that the goods are detected during the search journey. It is preferably provided that the method comprises determining at least one further parameter of the goods, such as the type of goods, and preferably identifying the goods. The parameter can, for example, comprise a class or type of goods and / or size, color, or weight values.It is preferably provided that the at least one further parameter of the item and preferably its identity is included in the check as to whether the item is located at a defined location. It is preferably provided that at least one defined location to which at least one item has been assigned is deleted if at least one item assigned to it or preferably all of the items assigned to it are removed from the location - preferably by at least one transport device. Depending on other conditions, a decision can be made not to carry out this deletion. For example, some locations in the database can be marked as non-deletable and in this case are not deleted. It is preferably provided that determining the pose of the item comprises i. determining the pose of the item in relation to the transport device and ii.the pose of the transport device is determined in relation to the inspection area and / or logistics area, and; iii. the pose of the goods is then determined in relation to the inspection area and / or logistics area, taking into account the two previously determined poses of points i and ii. This enables particularly simple determination of the pose of the goods in relation to the inspection area and / or logistics area. Points i and ii can occur simultaneously or one after the other in any order. Preferably, the perception and / or determination of the pose of the at least one goods is carried out at least partially by at least one TOF sensor, LIDAR sensor, camera, stereo camera, ultrasonic sensor, and / or radar sensor of the sensor arrangement. Perception and determination methods are already known in the prior art, for example from EP 3961 259 A1 or EP 3497672 A1.Preferably, the pose data of a defined location are corrected according to the pose of the item assigned to it if the check reveals that the deviation between the determined pose of the detected item and the pose data of the defined location assigned to it is above at least one defined threshold value. This enables a reaction to changes in the pose of the items. Preferably, the pose of the transport device is determined, and particularly preferably recorded, at least during its movement, continuously or at predetermined time intervals. Preferably, a software module on board the transport device interprets the raw sensor data to classify and / or identify specific items. Preferably, a software module on board the transport device determines the position and orientation of the items in relation to the transport device (pose determination).In principle, any sensor system that can provide data for object detection and pose determination can be used for this concept (mono / stereo camera, lidar, ultrasound, radar, TOF Time Of Flight, etc.). The software module can then evaluate this data and detect relevant objects or record their pose as stably as possible in 6 degrees of freedom. The analysis of the sensor data and processing of the detections can be carried out using conventional algorithms or AI methods (e.g. neural networks). With both approaches (different sensors and processing steps), sufficiently accurate and stable estimates for the pose of certain goods relative to the sensor system on the transport device could be determined. The detected goods are then transformed, filtered, numbered, tracked, and transferred to the map coordinate system of the transport device. Places can be assigned in various ways (e.g.User interface or via API requests via the API interface "Application Programming Interface"). This enables the software module to dynamically generate recording positions in order to autonomously process unstructured scenarios without precisely predefined workstations. Dynamic locations are sometimes referred to below. These are defined locations that are created and saved according to the inventive method through point d (creating at least one defined location and saving the created location). Example of a functional principle • Configuration and process definition o A specific area within the logistics area is predefined for dynamic station handling (checking area). This could be possible, for example, via specially defined areas / regions on the map. o Certain types of goods are selected that the system should recognize and locate in this area.oA workflow is defined which can include the following steps: Positioning at one or more search positions with a view of the search area Optional, as this could also happen while driving past Waiting time for switching between several search positions or aborting the workflow Pick operation for at least one of the recognized dynamic locations to pick up an item Transport to any target location or a group of locations Automatic deletion of the dynamic location oPickup poses are automatically created and updated by the logistics system as soon as dynamic locations are created or deleted. • Einsatzo The transport device moves to a search position or along a search route. During movement, the position of the transport device is continuously queried on the map and saved with a timestamp. o The sensor system scans the search area and provides the position and orientation of all detected goods relative to the sensor system. The pose of the measurement is transformed into the robot coordinate system. The timestamp of the measurement is used to determine the orientation and position of the transport device in the map coordinate system at that time and to transform the measurements accordingly into the map coordinate system. The detections are filtered based on the boundaries of the verification area, confidence, height above ground, and position in space. The positions of previously recorded and new goods are compared, and corrections are made if necessary.If the distance is too large or if no dynamic location exists in this check area, a new one is created. oIf a dynamic location is created, it is entered in the station list of the database. The occupancy of the location is also entered oIf an interaction is part of the workflow, the mobile robot starts by traveling to the location. An action is carried out at the location (e.g. picking up a good). If defined in the workflow, the location currently being processed is automatically removed from the database. Otherwise, an order can optionally be created within the system. Will be carried out by another transport device or the same one at a later time. If no order definition is required, the iteration of the process or workflow is completed. oIf the workflow provides for multiple runs, the process starts again after the order has been successfully processed.Preferably, after checking whether the item is located at a defined location, a further check is carried out to determine whether the item is assigned to the at least one defined location in which it is located, and that the item is assigned to the at least one defined location in which it is located if it has not yet been assigned to the location. This can be the case, for example, if the item was placed in the defined location in another way, for example manually. In this way, the database can be regularly updated. The invention will be explained in more detail below with reference to embodiments of the invention in the figures, in which: Fig. 1 is a flowchart of software for a transport device that uses the method according to the invention; Fig. 2 is a flowchart for the detection and determination of the item from Fig. 1; Fig.3 shows a schematic top view of a logistics area during a search run of a transport device while carrying out a method according to the invention without any inspection areas shown; Fig. 3a and Fig. 3b show a schematic top view of a logistics area during a search run of a transport device while carrying out a method according to the invention with inspection areas shown; Fig. 4a, 4b and 4c show a schematic top view of a logistics area during the organization of goods with a transport device in chronological sequence. Fig. 1 shows how the logic of a transport device can utilize the invention. First, after starting, a search run is carried out and the sensor arrangement is activated to search for goods (Block 1). In Block 2, a query is made as to whether a good has been detected. If so, the sensor data is transformed and filtered in Block 3. If not, the program jumps to Block 12.After transformation and filtering, block 4 determines the pose based on the transformed and filtered sensor data, checks it, and creates and saves it in the database, including determining a pickup pose, according to the invention. Block 5 then queries whether an interaction is defined. If so, block 6 moves to the pickup pose and aligns the transport device according to the pickup pose. If not, block 7 queries whether an order should be created. This is created in block 8 if necessary, but the program skips to block 12 in any case. After moving to the pickup pose, block 9 picks up the item from the defined location. Block 10 then queries whether the defined location should be deleted from the database based on at least one parameter of the defined location. If so, the location is deleted from the database in block 11.In any case, block 12 asks whether the search should be continued. If so, the process is restarted; otherwise, it is terminated. Fig. 2 shows details of the transformation and filtering from Fig. 1. First, the sensor data is transformed into the coordinate system of the transport device (block 15), which determines the relative pose of the goods to the transport device. Then, in block 16, the transformation into the coordinate system of the logistics area takes place by jointly transforming the transformed sensor data and the pose data of the transport device at the time the sensor data was recorded. Subsequently, in block 17, the data is filtered based on position, orientation, and confidence. Fig. 3, 3a, and 3b schematically explain an embodiment of the method according to the invention while the assignment process is being carried out.3 and 3a show a logistics area 1, which represents a storage room. In this storage room, one part has been defined as a two-part inspection area 2 (represented by a dashed line border). For better clarity, the inspection area 2 is not shown in Fig. 3. The transport device 3 is shown in Fig. 3 and 3a during a search run, while it moves along a transport path 4. During its movement, it has activated its sensor arrangement and thus checks the areas of the inspection area 2 that it passes through during the journey. Several defined locations 5 are already present, some of which are occupied by goods 6. The defined locations 5 have recording poses 7, represented by a point.Before the item can be picked up by the fork / lifting fork of the transport device 3, it must move to the position of the point, align the fork / lifting fork in the direction of the location 5 and move under the item 6 from the corresponding direction. In Fig. 3a, the items 6 have already been placed on defined locations 5, but have not yet been assigned to the defined locations 5 in the database. All defined locations 5 may be marked as free. An item 6 in the bottom left is placed in the inspection area 2, but not in a defined location 5. The transport device 3 detects the items 6 and determines their pose as well as the type of item 6. In Fig. 3a, two types of items 6 are visible, which differ in their size. The transport device 3 then checks whether each item 6 is in at least one defined location 5.In the case of item 6 at the bottom left, this is not the case, so a newly defined location 5a is created whose pose corresponds to this item 6 plus an expansion area and a safety area. Item 6 is assigned to location 5a, and the new location 5a is saved. This location 5a is thus saved in the database as location 5a occupied by item 6 of type A. The other items 6 are already partially located in defined locations 5, but they have not yet been assigned to them. During the check, it is therefore recognized that they are located in locations 5. Further checking, however, reveals that no assignment has been made. The respective item 6 with a respective type A or B is assigned to the respective locations 5. These locations 5 are thus saved in the database as location 5a occupied by item 6 of type A or as location 5b occupied by item 6 of type B.The expansion area arises from the fact that in this warehouse each defined location has a minimum size, due to the design of the transport device 3. The safety area is arranged around the remaining location 5 and serves to compensate for measurement or positioning errors. Fig. 3b shows the now updated defined locations 5, 5a and 5b. Figures 4a to 4c show a similar embodiment of the method according to the invention in chronological sequence. Again, the logistics area 1 is a warehouse, with an inspection area 2 defined. In this area, goods 6 of types A and B are randomly placed manually so that they can be automatically picked up and stored / rearranged by the transport device 3. In Fig. 4a, the transport device 3 moves into the inspection area 2 and carries out the allocation process explained.None of the goods are stored at a defined location 5, therefore a total of five new defined locations 5 are created, the goods 6 are assigned with their corresponding type A / B, and saved. Fig. 4b shows these updated defined locations 5, 5a, 5b. A transport device 3 or an interaction on the transport device 3 (not shown) can then process the process in which a good 6 is picked up and placed at a predefined location 5 in the warehouse. The locations 5 created during the allocation process are deleted as soon as they are free again. This is shown in Fig. 4c. The interactions discussed can be carried out by the transport device 3 that executes the allocation process or by another transport device 3.It can be provided that the transport device 3 that executes the assignment process transmits orders for interactions to other transport devices 3. This can also be provided in other embodiments according to the invention.

Claims

1. Method for organizing goods (6) in a logistics area (1) with at least one driverless transport device (3), wherein at least one inspection area (2) is defined and the method comprises an allocation method which has the following steps: a. Perception of at least one item (6) located in the inspection area (2) by at least one sensor arrangement of the transport device (3); b. Determination of the pose of the item (6); c. Checking whether the item (6) is located at a defined location (5, 5a, 5b), wherein the check comprises comparing at least the pose of the item with pose data from defined locations, wherein the defined locations, including their pose data and occupancy data, are stored in a database; d.Creating at least one defined location (5, 5a, 5b) if the check has determined that the item (6) is not located at at least one defined location (5, 5a, 5b), wherein the created defined location (5, 5a, 5b) includes the pose of the item, assigning the item (6) to the created at least one defined location (5, 5a, 5b), and storing the created location (5, 5a, 5b) in the database.

2. Method according to claim 1, characterized in that, preferably if the check could not assign the item (6) to at least one defined location (5, 5a, 5b), at least one pickup pose of an item is determined, which is assumed by a transport device (3) during or before picking up the item in question, and that this pickup pose is preferably stored in the database. 3.Method according to claim 1 or 2, characterized in that the detection of the item (6) occurs while the transport device (3) is traveling.

4. Method according to one of claims 1 to 3, characterized in that at least one transport device (3) performs at least one search run while moving in the inspection area (2) or along the inspection area (2), and that the detection of the item (6) occurs during the search run.

5. Method according to one of claims 1 to 4, characterized in that the method comprises determining at least one further parameter of the item, such as the type of item, and preferably identifying the item (6).

6. Method according to one of claims 1 to 5, characterized in that the at least one further parameter of the item, and preferably its identity, is included in the check to determine whether the item (6) is located at a defined location (5, 5a, 5b).Method according to one of claims 1 to 6, characterized in that at least one defined location (5, 5a, 5b), to which at least one item (6) has been assigned, is deleted if at least one item (6) assigned to it, or preferably all of the items (6) assigned to it, are removed from the location (5, 5a, 5b) - preferably by at least one transport device (3).

8. Method according to one of claims 1 to 7, characterized in that determining the pose of the item comprises determining the position of the item in the inspection area (2) and / or logistics area (1) and / or determining the location of the item in relation to the inspection area (2) and / or logistics area (1).

9. Method according to one of claims 1 to 8, characterized in that determining the pose of the item comprises i. determining the pose of the item in relation to the transport device (3) and; ii.the pose of the transport device (3) is determined in relation to the inspection area (2) and / or logistics area (1) and; iii. then the pose of the item is determined in relation to the inspection area (2) and / or logistics area (1) taking into account the two previously determined poses of points i and ii.

10. Method according to one of claims 1 to 9, characterized in that the perception and / or determination of the pose of the at least one item of goods is carried out at least partially by at least one TOF sensor, LIDAR sensor, camera, stereo camera, ultrasonic sensor and / or radar sensor of the sensor arrangement.

11. Method according to one of claims 1 to 10, characterized in that the pose data of a defined location (5, 5a, 5b) are determined in accordance with the. The pose of the item assigned to it can be corrected if the check reveals that the deviation of the determined pose of the detected item from the pose data of the defined location (5, 5a, 5b) assigned to it is above at least one defined threshold value.

12. Method according to one of claims 1 to 11, characterized in that the position of the transport device (3) is determined continuously or at predetermined time intervals, at least during its movement, and particularly preferably recorded. 2025 03 18 MT

Images