Method, goods allocation device, and system for packing goods

Abstract

The disclosure relates to a computer-implemented method for packing one or more goods in a, more particularly predefined, package and / or in a corresponding goods receptacle, more particularly a goods submission opening, of a packing station, comprising: producing, on the basis of image data that represent a plurality of goods having at least two, more particularly different or identical, properties, an individualized representation of the plurality of goods, determining, more particularly identifying, on the basis of the individualized representation, an individual good from the plurality of goods, and / or at least one property associated with the individual good; and outputting a control signal on the basis of the at least one determined property. The disclosure also relates to a computer-implemented method for allocating and / or packing one or more goods by means of a determined instance from a plurality of packing stations, comprising: producing, on the basis of image data that represent a plurality of goods having at least two, more particularly different or identical, properties, an individualized representation of the plurality of goods, determining, more particularly identifying, on the basis of the individualized representation, an individual good from the plurality of goods, and / or at least one property associated with the individual good; and outputting a control signal on the basis of the at least one determined property. The disclosure additionally relates to a system for packing goods and to an applicable goods allocation device that more particularly implements the methods described herein. Moreover, a computer program product and a computer-readable storage medium are provided.

Description

[0001] Method, goods allocation device and system for packaging goods

[0002] The present invention relates to the packaging of goods, and in particular to a method and a system for packaging goods, as well as a corresponding goods allocation device that particularly implements the inventive method. Furthermore, a computer-implemented method, a computer program product, and a computer-readable storage medium are provided.

[0003] Conventional processes and systems for packaging goods usually require a large number of workers, for example to sort and / or classify a large number of incoming goods, for example to assign them an identifier so that the goods can then be assigned to a specific customer for packaging. However, such processes and systems are not only extremely time-consuming, for example when large quantities of goods are involved, but also cost-intensive. Employing a large number of workers for manual packing can significantly increase operating expenses and have a negative impact on profitability, particularly when large quantities of goods need to be packed. Furthermore, the throughput rate and / or performance of such processes and systems is limited, in particular by the performance of the workforce.

[0004] Another problem with manual processes is their limited flexibility. Employees require specialized training to identify, sort, and package or route different types of goods. Changes to the product range or the introduction of new items may require retraining, requiring additional time and / or resources.

[0005] Furthermore, manual packaging processes reach their limits as demand increases. Scalability is challenging, as it can be challenging to find enough workers to keep up with the increased production rate, especially during peak periods.

[0006] There are also approaches to methods and systems for packaging goods that are at least partially automated, for example, through the use of robots and / or goods distribution systems, e.g., conveyor belts. However, these methods and systems still require employees, as the automated detection, identification, selection, picking, and / or grasping of a single item, e.g., from the conveyor belt and / or from a box or crate containing multiple goods or products, e.g., in a disorganized / unsorted manner, presents a particular challenge.

[0007] It is an object of the present invention to overcome the disadvantages of the known prior art and in particular to provide an improved method or an improved system for packaging goods, as well as a corresponding goods allocation device.

[0008] The problem is solved by the features of the independent claims. The dependent claims describe preferred embodiments. Further aspects, advantages, and features emerge from the dependent claims, the description, and the accompanying drawings.

[0009] The invention relates to a computer-implemented method for packaging one or more goods in a, in particular predefined, packaging and / or in a corresponding goods receptacle, in particular a goods input opening, comprising:

[0010] - generating, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualised representation of the plurality of goods,

[0011] - Determine, in particular identify, on the basis of the individualized representation,

[0012] -- an individual product from the plurality of products, and / or

[0013] -- of at least one characteristic associated with the individual product; and,

[0014] - Outputting a control signal based on the at least one specific property.

[0015] The invention relates to a computer-implemented method, in particular according to one or more of the aspects and / or embodiments described herein, for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations, comprising: - generating, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualized representation of the plurality of goods,

[0016] - Determine, in particular identify, on the basis of the individualized representation,

[0017] -- an individual product from the plurality of products, and / or

[0018] -- of at least one characteristic associated with the individual product; and,

[0019] - Outputting a control signal based on the at least one specific property.

[0020] The invention also relates to a goods allocation device according to the invention, in particular a robot, in particular a pick-and-place robot, for packaging one or more goods, which implements the method according to the invention according to one or more of the aspects and / or embodiments described herein.

[0021] The invention relates to a system for packaging one or more goods, which comprises the goods allocation device according to the invention and at least one image recording device, in particular with at least one camera, for obtaining the image data.

[0022] The invention relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to the invention according to one or more of the aspects and / or embodiments described herein.

[0023] The invention relates to a computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to carry out the method according to the invention according to one or more of the aspects and / or embodiments described herein.

[0024] It should be understood that the embodiments described in the claims and in the description in connection with the system and / or the goods allocation device can preferably also be used in the method according to the invention as such, and vice versa. When reference is made herein to the method, the goods allocation device, or the system, this refers to the method according to the invention, the goods allocation device according to the invention, or the system according to the invention. "Packaging" can mean inserting, introducing, inserting, inserting, pushing in, placing, and / or arranging a product in and / or on a packaging of a specific type and / or type, e.g., a container, a pouch, a bag, a carton and / or paper, a film, a can, and / or packaging made of glass, metal, a plastic, e.g., plastic, and / or a sleeve, etc.be understood, for example, to protect, store, transport and / or contain a product and / or goods.

[0025] "Packaging" can therefore refer, for example, to inserting, introducing, inserting, inserting, pushing, placing, and / or arranging a product into a final package, which can, for example, be a carton, box, mailing bag, envelope, and / or parcel formed and / or folded from a packaging raw material. These terms can be used interchangeably herein. The product can, for example, be inserted into at least one product receptacle formed in the final packaging, for example, a product entry opening. This can then be closed, for example, to ship the package.Alternatively or additionally, “packaging” can also refer to laying down, depositing, placing, aligning and / or arranging the goods i) at a specific position and / or location on the packaging raw material itself, for example before the final packaging is formed and / or folded therefrom. The laying down, depositing, placing, aligning and / or arranging of the goods can take place, for example, on an inclined plane, in particular at least in sections, as a result of which the goods can slip or slide into an opening of the packaging machine; and / or ii) for example on a tray (for example a storage tray with or without side limitations, e.g. an open surface) on which the goods can be placed, whereby the packaging can be carried out by, for example, folding a box, e.g., around the goods. The same applies to the other packaging types and / or kinds mentioned herein.

[0026] Packaging can be used to protect the product from damage, contamination and / or other influences (e.g. temperature), for example during transport, and / or to avoid information about the product and / or the goods packed therein. The state when the package is not yet completely, or only partially, formed from the material can be referred to as unformed packaging, e.g. packaging raw material. At this stage, the packaging can often consist of a flat, unprinted material such as paper, cardboard, plastic film or a similar material. The packaging, in particular a flat one, can then be folded, shaped, formed and / or optionally printed according to various specifications to obtain the final packaging, e.g. the final packaging. The final packaging can comprise at least one goods receptacle, in particular a goods input opening for one or more goods and / or products, e.g.for bringing in goods.

[0027] “Predefined” packaging can be understood as end packaging and / or raw packaging material with a specific, defined and / or predetermined dimension, for example length, width, height, depth, thickness, in particular (packaging) wall thickness, a specific material and / or a specific material property, a specific shape, a specific weight, a specific surface finish and / or a specific robustness, strength and / or stability. Predefined packaging can also mean specific and / or predetermined customer-specific information, e.g. whether the package is an individual order or a collective order, e.g. several goods in one package or a partial order, e.g. a division of multi-part goods into different packages. Predefined packaging can, for example, be one or more containers orContainer(s), such as a V-mail, a bag, a parcel, a package, a box, a bundle, a vessel, a can, a carton and / or a container. Example materials for a predefined packaging can include one or more of: cardboard, corrugated cardboard, plastic, e.g. PET, PVC, PE, PP, plastic, wood, glass, metals, e.g. steel, aluminum, paper, jute, cotton. The predefined packaging can, e.g. depending on the shape and / or dimensions of the packaging and / or the goods to be packaged therein, comprise a cushioning and / or filling material. The type, quantity and / or size of the cushioning and / or filling material can, for example, be determined depending on the goods to be packaged in the predefined packaging, in particular on the basis of and / or depending on at least one property of the goods, e.g. weight, fragility, etc. The type of predefined packaging, e.g. bag instead of carton or vice versa, or a specific, e.g.more stable / larger carton, etc., can be determined, for example, depending on the goods to be packaged in the predefined packaging, in particular on the basis of and / or depending on at least one property of the goods, e.g. weight, fragility, etc.

[0028] The final packaging can, for example after it has been formed and / or folded, comprise at least one goods receptacle, in particular a goods input opening, into which the goods can be inserted. Alternatively, the final packaging can comprise at least one goods receptacle, in particular a goods input opening, while it is being created. As already mentioned, the final packaging can be predefined, e.g. its external dimensions can be predetermined. Example external dimensions can be in the range of approximately 35cm x 25cm x 2cm to approximately 120cm x 60cm x 60cm for the final packaging. This information on the packaging size is only to be seen as an example, although it corresponds to the typical sizes of most shipments of goods. The external dimensions of the final packaging can, for example, correspond to and / or be adapted to a goods receptacle, in particular a goods input opening, e.g.in terms of size and / or shape, for example, with a goods receptacle, in particular a goods input opening of a packaging station, e.g., a goods insertion slot of the packaging station. Other types of goods receptacles, in particular goods input openings, are also conceivable and depend on the packaging used or the packaging station.

[0029] The goods can therefore, for example, be packaged (itself) directly in a final packaging or can also be transferred directly to a packaging station and, for example, inserted, introduced, inserted, placed, arranged and / or plugged into its goods receiving device, in particular its goods input opening.

[0030] The packaging, especially predefined packaging, can

[0031] - a packaging raw material, in particular a substantially flat one; and / or

[0032] - comprise a final packaging, in particular a package, formed from the packaging raw material, in particular folded. The final packaging can comprise at least one goods receptacle, in particular a goods input opening.

[0033] According to the invention, image data are generated and / or obtained which represent a plurality of goods with at least two, in particular different or identical, properties.

[0034] For example, the plurality of goods may be located on a conveyor belt and / or in a container, box, and / or carton. The terms container, box, and / or carton may be used interchangeably herein.

[0035] The plurality of goods, in particular at least two, can either i) have the same properties, e.g. if the container exclusively contains, e.g., razors, scissors, calculators, ballpoint pens, contact lenses, etc., e.g. of the same type, for example razors of the same brand, size, type, etc.; or ii) at least two different properties, e.g., a razor and scissors; or two razors of different types, e.g., different brands, sizes, type, etc. The plurality of goods (identical or different) can be arranged and / or distributed in the container, in particular chaotically, in particular on top of one another, next to one another, underneath one another, overlapping, concealing one another, and having a, in particular random, alignment, position, location, and / or orientation in the container or box. Such a container can also be referred to as a "chaotic box." The goods therein can, for example,be referred to as bulk material.

[0036] According to the invention, the image data can be obtained and / or captured by at least one image recording device, for example, at least one camera. In principle, however, other types of detectors or image recording devices can also be used, such as laser-based detectors, in particular "Light Detection and Ranging" (LiDAR), ultrasound, ToF, laser scanners, radar sensors, etc. It should be noted that several, e.g., at least two or more, image recording devices can also be used. Preferably, however, several cameras can be used, e.g., to capture and / or record the contents of a, for example, chaotic, box, e.g., from different (viewing) angles and / or directions. The camera can be a 2D and / or a 3D camera, which can also capture spatial depth, for example. Infrared cameras are also possible.The image data can comprise a two-dimensional representation (2D) and / or three-dimensional representation (3D), which, for example, comprise one or more pixels and / or one or more images of the, in particular chaotic, box, in particular of its contents, in particular of the plurality of goods. The type of camera to be used, e.g. 2D or 3D, can depend on the number and / or type of AI systems to be used, as described herein. For example, a 2D camera may be sufficient if, for example, at least two AI systems are used, e.g. a first AI system, e.g. an LLM, and a second AI system, e.g. a CNN; however, this is only a non-limiting example, i.e. there are also other possibilities or combinations.

[0037] According to the invention, an individualized representation of the plurality of goods can be generated, in particular on the basis of the image data.

[0038] An "individualized representation" can be understood as a (2D and / or 3D) representation, e.g., of the plurality of goods, e.g., within the chaotic box, in which a, e.g., first, good is recognizable, identifiable, and / or detectable separately, individualized, segmented, rasterized, isolated, and / or detached from a, e.g., second, particularly identical or different, good. The generation of the individualized representation can therefore, for example, be referred to as object, product, and / or goods recognition or segmentation and / or can be carried out automatically. The individualized representation can, for example, comprise a three-dimensional representation (3D), e.g., including a spatial depth that includes, for example, one or more pixels and / or one or more images, e.g., of the chaotic box, in particular its contents, in particular of the plurality of goods. For example,the goods are presented individually and separately from one another, including in terms of depth, e.g. when the goods are placed on top of one another.

[0039] The generation of the individualized representation may comprise processing the image data by means of at least one artificial intelligence (AI) system.

[0040] An AI system (artificial intelligence system) can refer to a technological application based at least on artificial intelligence and / or a system equipped with AI to automate and / or support human-like skills and tasks. According to the invention, the at least one AI system can be used, e.g., for image, object, or product recognition and / or segmentation, e.g., in a chaotic box, for picking up and / or forwarding the object or product, and / or for packaging the object or product.

[0041] At least one AI system or several, e.g. at least two, e.g. different or identical, AI systems can be used for the present invention. The at least two AI systems can exist and / or be operated or work independently of one another or can at least partially influence one another. The at least two AI systems can, for example, work in parallel and / or alongside one another (e.g. independently of one another or dependently of one another) or one after the other, sequentially and / or in mutual exchange with one another, so that, for example, an output of a first AI system can serve as input for or be provided for a second AI system, or vice versa. In the event that at least two AI systems are used, it is conceivable that a first AI system and a second AI system communicate with one another or are at least connected to one another, for example communicatively.Additional KL systems can be used, which in turn can be communicatively connected to the first / second KL system.

[0042] The first AI system and the second AI system can, for example, receive essentially the same / identical input, for example the image data described herein and / or the individualized representation. Alternatively or additionally, the first AI system and the second AI system can share data and / or be connected to, for example, the same, image recording device or the same sensor, for example at least communicatively. However, the input can also be different. For example, it is possible for one of the first or second AI system to receive the image data described herein and / or the individualized representation as input, and the other of the first or second AI system to receive the corresponding output after processing the image data described herein and / or the individualized representation.

[0043] The at least one AI system (or several) can, for example, be provided locally, for example at or at least in the immediate vicinity of the goods allocation device, in particular the packaging machine, or the packaging location A place and / or, preferably directly, coupled to or connected to a control system of the goods allocation device, or at least in the immediate vicinity thereof, e.g., a control and / or monitoring station or a computer with appropriate performance. For example, a preferably communicative / data-based connection, e.g., a data line, can be provided between the AI ​​system and a packaging machine. In the case of an LLM, as described herein by way of example, a local solution would be, for example, a version of the "Llama" language model from Meta. Alternatively, the at least one AI system (or several) can, for example, be provided centrally, e.g.,be provided at another location, on / in / via a remote cloud or a server, for example at a provider of the AI ​​system, especially if the AI ​​system is received as a third-party service. A combination of local / central is also conceivable. For example, a first AI system can be provided locally and a second AI system can be provided centrally, e.g. at the provider's, or vice versa. When using multiple goods allocation devices or packaging stations, e.g. within a packaging and / or logistics department of a company, it can be useful if one or more AI systems can communicate with each other, e.g. are communicatively connected, so that, for example, all local data can be combined to form a large data treasure and / or stored locally or in a cloud.

[0044] The first AI system may differ from the second AI system. For example, one of the first or second AI systems may be a so-called "strong" AI system, and another of the first or second AI systems may be a so-called "weak" AI system. The difference between "strong" AI and "weak" AI may lie in the goals, properties, and / or capabilities of the artificial intelligence. A "weak" AI system may be designed to solve specific application problems or master concrete tasks in individual domains. Such systems may simulate intelligent behavior (but do not exhibit general intelligence themselves) to address a specific problem, without, for example, developing a deeper understanding and / or awareness of their actions.Examples of weak AI systems are voice assistants (such as Siri or Alexa), image recognition systems, chess programs, or recommendation systems (such as those used by Netflix or Amazon). These systems can be highly specialized and / or powerful in a narrowly defined area. Weak AI can mimic intelligent behavior, but may not be able to fully replicate human thought and / or consciousness. Weak AI can follow certain algorithms and / or models to complete tasks without, for example, developing an understanding of the underlying problems and / or context. However, weak AI can also have the ability to learn, for example through machine learning, and / or improve its behavior. Weak AI systems are well suited to solving specific, clear, and / or well-defined (specialized) tasks and / or can be very powerful in specialized application areas.

[0045] For the present invention, the AI ​​system and / or one of the first or second AI systems can, for example, comprise at least one weak AI or at least one weak AI system, which, as a non-limiting example, deals with determining coordinates, in particular spatial coordinates, e.g. x-coordinate, y-coordinate and / or z-coordinate, for a receiving point or pick-up point of the goods, e.g. by means of at least one gripper of the goods allocation device and / or defines a pick-up strategy. The pick-point strategy can include defining the pick point in the area center of gravity or the mass center of gravity, or sections / regions / areas with, for example, particularly grippy / absorbent surfaces and / or a type of suction device / gripper, as well as a possible approach movement, approach route, approach path and / or a rotation, alignment and / or orientation of the gripper. The pick-up point can be an optimal location orbe an optimal area, point and / or section at which the goods can be picked up, e.g., gripped. The pick-up point can, for example, correspond to a center of gravity, in particular a center of gravity of the goods. The learning loop can essentially be limited to the exploitation of probabilities and / or weightings of success or failure when the goods are transported from the pick point to the packaging machine. A success can, for example, be that the picked / picked goods could be transported or packaged for packaging. A failure can, for example, be that the picked / picked goods could not be transported or packaged for packaging, in particular not without defects. A weak class can, for example, quantitatively describe at least one property of the goods or of the individual goods, such as, for example,Determine the approximate dimensions based on the outlines and / or, as mentioned above as an example, determine the pick point, although other applications would also be conceivable / possible.

[0046] A "strong" AI or a "strong" AI system is currently still a hypothetical concept and can possess cognitive abilities comparable to and / or surpassing human reasoning. A strong AI can not only perform specific tasks (cf. weak AI), but also have a general understanding of relationships and / or problems, similar to the human mind. For example, a strong AI can independently complete difficult tasks and / or think, plan, or solve problems in a way that is at or above human level, for example, on par with humans who can take on work to complete difficult tasks. In the future, however, it is conceivable to use a strong AI for invention. In theory, a strong AI would be capable of thinking, learning, and / or making decisions autonomously that are meaningful beyond a specific context.A strong AI may exhibit the ability for self-reflection, awareness, and / or general cognitive abilities such as logical reasoning, planning, creativity, and / or abstraction. For the present invention, a strong AI could qualitatively recognize at least one property of the product or individual product, for example, weight, fragility, etc., as well as identify the product itself. The use of a strong AI could reduce the requirements for the image capture device (e.g., 2D instead of 3D) and / or the number of image capture devices.

[0047] A learning loop of the "weak" AI mentioned above can, for example, be limited within certain limits to exploiting probabilities and / or weightings of success or failure. A "weak" AI can be understood as one that performs its functions more or less "strongly" or "weakly" by taking probabilities and possibly other parameters into account, or whose decision-making power is more dependent on the available weightings. In principle, the same can apply to the representation of "strong" AI or "strong" AI systems, which, depending on their development, provide a greater or lesser degree of intelligence, learning capacity, learning outcome, accuracy, and / or speed within a certain range.

[0048] It is possible for the first AI system to differ from the second AI system in terms of a type of AI system. In particular, the first AI system may be a language model, in particular a Large Language Model (LLM), and the second AI system may be a Convolutional Neural Network (CNN), as described herein. However, it is also possible for the first AI system and the second AI system to be essentially identical, e.g. a CNN or an LLM. Other combinations are also conceivable, although it should be clear that more than two AI systems may also be used, e.g., a third, fourth, fifth, sixth, etc. The first and second AI systems may be located at different locations.

[0049] The first AI system can, for example, be located on a goods allocation device, for example, at a goods receiving end point, attached, for example, to a gripper, in particular the goods allocation device, for example, a pick-and-place robot. The second AI system can, for example, be included or integrated in a higher-level system of the packaging company, but this is only an example. For example, the first AI system can perform an estimation and / or calculation before and / or during the picking of the goods from the chaotic box, for example, in order to estimate or pre-estimate at least one property of the individual goods. For example, the first AI system can use the image data (and / or the individualized representation) to, for example,In a first step or pre-processing step, i) recognize at least one text, symbol, trademark, sign, color, pattern, imprint, inscription, graphic design, number, special character, marking, note, sticker, and / or inscription on the product or an individual product. For example, the first class system could recognize that the word "Edding" is written on a pen in the particularly chaotic box and thereby infer at least one property, e.g., size, length, thickness, surface texture, etc. ii) recognize at least one barcode, matrix code, barcode, and / or QR code, if necessary.to retrieve this information and, for example, to derive weight information of the goods, in particular of the individual goods; iii) to recognise at least one shape, outline, contour, geometry, structure and / or dimension of the at least one good, in particular of the individual good, for example that the good is a pair of scissors; etc.

[0050] In order to improve the output of the first AI system (and / or second AI system), at least one prompt can be used which can provide the AI ​​system with instructions, in particular textual ones, e.g. which procedure is advantageous when analyzing and / or creating the image data described herein and / or the individualized representation.

[0051] The result of this first step or of the pre-processing step by the first AI system can be an output, in particular a first output, which is indicative of the at least one property of the product or of the individual product. This first output can then be fed back to the first AI system, e.g. as input, whereupon the latter can, e.g. in a second step or a processing step, process / analyze the individual product, e.g. with regard to the at least one property assigned to the individual product. The result of this second step or of the processing step can be the control signal described herein. In the case just mentioned, for example, only at least one AI system would be required which operates / processes step by step.

[0052] If at least two AI systems are used, the first output of the first AI system can be made available to the second AI system, e.g. as input. In this case, the first AI system could be an LLM, for example, which passes the first output on to the second AI system, e.g. a CNN. The findings of the first and second AI systems can then be summarized and, e.g. influence the control signal or be integrated into it. Optionally, the findings can be compared with learned, stored, trained and / or predetermined data or experience, results and / or outcomes in order to increase efficiency / accuracy. For example, the findings of the one or more AI systems can include one or more of: a center of mass, a pick point or pick-up point, a pick tool, the type of packaging or packing location, the handling orHow the article should be inserted into the packaging or into the packaging machine, etc. Alternatively or additionally, action information for the goods allocation device can be generated or included, such as start and / or target positions, gripping parameters, speed parameters, etc.

[0053] The first output can, for example, give the CNN an indication of which product has been detected, e.g. that it is a pen, in particular a permanent marker, which has a specific dimension or a specific weight and includes the optimal, in particular spatial, pick-up coordinates (x, y, and / or z), whereby alternatively or additionally a pick-up strategy for the pen, for example, can also be defined. This procedure can reduce errors and save time, for example when processing the data by the second AI system, since the first output can be used as additional information. The second AI system can then generate the control signal, e.g. as a second output.

[0054] The at least one or more AI systems can, for example, cooperate, collaborate, and / or complement each other (e.g., "human-AI collaboration") with a human / employee, or an operator, or a user, e.g., of the packaging machine or a packaging person, in order to, for example, improve visual recognition of the goods or individual goods in the, particularly chaotic, box, for example, in cases where the one or more AI systems reach their limits. The human can serve as a kind of "intelligence amplifier" by intervening, either directly or at least indirectly, and / or confirming or correcting / adapting the AI's suggestions. This intervention can be viewed as feedback or a feedback signal that the AI ​​can use, for example, to improve its performance and / or accuracy.

[0055] As an example, the employee can be informed of an error or a problem, e.g. by means of an indication, in particular a display, or e.g. by laser marking, in particular a laser pointer, e.g. on the product itself. The employee can then intervene and corresponding findings / results from the process (or parts thereof) can be used as feedback or feedback signal. It would be conceivable, for example, for the employee to intervene, in particular directly, in the packaging process. For example, confirmation, approval and / or at least one input may be required so that, for example, the packaging process can be continued, in particular correctly. It is conceivable, for example, that the product, or the individual product, is arranged, positioned and / or aligned in an unfavorable manner, particularly for visual recognition. The employee can then rearrange, position and / or align the product so that the new position, e.g.can be (better) identified using the image capture device. This approach can, for example, prevent or unravel a jumble of different goods, e.g., in a chaotic box. In this scenario, the employee can contribute to solving the problem (incorrect / inaccurate positioning).

[0056] Alternatively or additionally, it is also conceivable that the at least one or more AI system(s) contribute to solving such problems (e.g., incorrect / inaccurate positioning of the goods) (e.g., a "mechanical solution"), for example, to improve visual recognition of the goods or individual goods in the, in particular, chaotic box. For example, the at least one or more AI system(s) can instruct, control, and / or monitor the goods allocation device to move, realign, and / or reposition the affected goods or individual goods, in particular at least slightly or sufficiently so that the image recording device can, for example, generate new image data.

[0057] Through such feedback or human intervention, the efficiency of automation processes can be optimized. Machine interventions, such as adjusting the position of goods, can complement or at least partially replace human assistance and demonstrate how such hybrid systems could shape future production processes.

[0058] An AI system can use algorithms, models, networks, and / or data to process information, recognize patterns and / or relationships, make decisions, and / or solve problems. An AI system can include a neural network, a deep learning system such as a convolutional neural network (CNN), a language model such as a large language model (LLM), and / or machine learning. An AI system can have a learning process in which it learns from provided training data and improves its performance over time. This can be done, for example, via supervised learning, unsupervised learning, and / or reinforcement learning.

[0059] Training data can, for example, correspond to the image data and / or the individualized representation and be provided to at least one AI system as input.

[0060] For example, one or more images of the, in particular chaotic, box, in particular its contents, in particular the plurality of goods, can be provided to the at least one AI system, wherein the image data can represent a plurality of goods with at least two, in particular different or identical, properties. On the basis of the image data, the at least one AI system can then generate one or more individualized representations which can be evaluated, e.g., by means of feedback. Possible feedback methods include, for example, user feedback, in which users can give the at least one AI system direct feedback on the generated individualized representation, e.g., confirmation that the individualized representation was generated correctly (e.g., all goods in the box were recognized); automatic feedback, for example, from the at least one AI system itself, e.g., based on orAnalysis of user interaction and / or behavior and / or if the at least one AI system notices, for example, that a product was not processed correctly, e.g. was not tangible and / or could not be inserted into packaging and / or a product receptacle, in particular a product input opening; and / or error detection, for example if the at least one AI system (itself) draws the user's attention to problematic, uncertain, difficult and / or critical data, e.g. if the at least one AI system recognizes products incorrectly, faultily, ambiguously, only partially and / or not at all and / or creates an incorrect or doubtful individualized representation.

[0061] Alternatively or additionally, one or more generated individualized representation(s) can also be provided to the at least one AI system, wherein the individualized representation can be the plurality of goods, e.g. within the chaotic box, in which one, e.g. first, item of goods can be recognized, identified and / or captured separately, individualized, segmented, rasterized, isolated and / or detached from a, e.g. second, in particular identical or different, item of goods. Based on the individualized representation, the at least one AI system can decide, e.g. how the item is to be picked up from the chaotic box, e.g. how it is to be grasped and / or how it is to be further processed, e.g. how it is to be packaged. This can also be evaluated, e.g. user feedback can be given, in which the user gives the at least one AI system, e.g. directly feedback on the picked up item, e.g.confirmation that the goods have been picked up correctly; automatic feedback, for example from the at least one AI system itself, e.g. based on or analysis of the interaction of a gripper and / or at least one sensor with the goods, e.g. if the at least one AI system detects that goods have not been processed correctly, e.g. were not graspable and / or could not be inserted into packaging and / or a goods holder, in particular a goods input opening; and / or error detection, for example if the at least one AI system draws the user’s attention to problematic, uncertain, difficult and / or critical data, e.g. if the at least one AI system detects goods incorrectly, faultily, ambiguously, only partially and / or not at all and / or does not know how to grip them, e.g. due to their shape.

[0062] As training data, alternatively or additionally, at least one property of the product, one or more reference values, data from one or more sensor devices and / or even the feedback itself can be provided (see below).

[0063] The "feedback" described herein can take place for one or more training data in the form of one or more of the following: user feedback, as described above, automated feedback, and / or error detection, as described above. In supervised learning, the model can be presented with training data, e.g., labeled data sets, consisting of input attributes (features) and the corresponding target values ​​(labels). The goal can then be to learn a function that maps the mapping of input attributes to the correct target values. The model can be iteratively adjusted during training to make the best predictions for new, e.g., previously unseen or otherwise used data. Examples of supervised learning are classification and regression problems. In a classification problem, inputs are to be assigned to specific classes (e.g., image recognition: razor or contact lens).In regression problems, continuous values ​​are to be predicted (e.g. house price prediction based on certain characteristics).

[0064] In contrast, unsupervised learning cannot use target values ​​or labels for the training data. Instead, it can aim to discover patterns, structures, and / or relationships in the data itself, e.g., autonomously and / or unsupervised, without the model having any prior information about how the data should be categorized or grouped. The model learns to group, segment, or reduce the data based on similarities or other features, e.g., exploratory. Examples of unsupervised learning include cluster analysis, dimensionality reduction, and association rules.

[0065] A neural network can be a model, particularly an algorithmic one, that is inspired by and / or modeled on the functioning of the human brain. The neural network can comprise a plurality or collection of interconnected neurons that can process and transmit information and / or recognize and / or learn patterns or relationships. By training on large amounts of data, a neural network can be capable of learning complex tasks such as image recognition, segmentation, language processing and / or decision-making. A special type of neural network, e.g. the CNN, can use specific layers, such as convolutional layers and / or pooling layers, to extract features and / or properties from images and to recognize patterns and / or relationships. Bounding boxes can also be used to obtain an individualized representation.For example, after the goods have been detected, e.g. in the image data, a frame, e.g. a bounding box, can be automatically placed around each individual object or each item of goods (e.g. detected in the chaotic box), which defines the position, location, alignment, orientation and / or extent of the goods in the image data. In particular, 2D and / or 3D bounding boxes that additionally take spatial depth into account are conceivable. The at least one AI system can also be trained with a large amount of training data (e.g. in advance and / or during operation), which includes, e.g., images of goods, e.g. in a chaotic box. In this way, the CNN can learn to identify features, relationships and / or patterns that are relevant or useful for the detection, detection and / or segmentation of the goods. A CNN can also be based on convolutional operations.These convolutional operations can enable effective capture of features in images, such as edges, textures, and / or shapes, that are useful for object recognition and discrimination. A CNN can also extract hierarchical features by using different layers of neurons to learn increasingly complex information. The at least one AI system described herein may, for example, be a CNN.

[0066] As already mentioned, the image data, e.g., of a chaotic box, and / or the individualized representation can be provided to one or more AI systems, for example, as input. It is also possible for the image data and / or the individualized representation to be provided to one or more AI systems as training data (see above).

[0067] If, for example, at least one AI system is used, it can be provided that the at least one AI system works step by step, e.g. in several steps, or evaluates the image data or the individualized representation and / or feeds back the output, e.g. for further analysis / processing, for example comprising at least a first step or preprocessing step and a second step or processing step. Alternatively, at least two AI systems can be used. For example, the first AI system can receive a first input (e.g. the image data described herein and / or the individualized representation) and, on the basis of this, generate a first output which can be provided to the second AI system as input, in particular a second input. The second AI system can generate a second output which, for example, can correspond to the control signal or can be integrated therein.

[0068] A language model or a large language model (LLM for short) can be a type of artificial intelligence (AI) that is configured to understand, generate, and / or respond to language. The AI ​​system or model can be based on a deep learning neural network architecture and preferably comprise a machine learning system, in particular a deep learning neural network. The AI ​​system or model can therefore have a learning process in which it learns from, for example, provided training data and improves its performance over time. An example of a language model or LLM is Chat-GPT, or GPT-4o (Generative Pre-trained Transformer), or similar advanced models. Other examples include BERT (Bidirectional Encoder Representations from Transformers), which can also be used to generate high-quality text.BERT is a language model based on the Transformer architecture that aims to improve bidirectional context processing.

[0069] Based on the individualized representation, an individual product can be determined from the plurality of products, in particular identified, and / or at least one property associated with the individual product. An "individual product" can refer to a product, e.g., a first product, that has been recognized, identified, and / or recorded separately, individualized, segmented, rasterized, isolated, and / or detached from a product, e.g., a second product (in particular, the same or different), e.g., from the plurality of products, e.g., in the chaotic box. The at least one property associated with the individual product can include a position, location, orientation, alignment, e.g., in the chaotic box, a surface quality, dimension, e.g., size, length, height, depth, thickness, shape, weight, stability, strength, hardness, and / or a material, as well as, e.g., a density of the individual product.Other additional properties can also be determined, in particular identified, for example a fragility or a degree of fragility, which can provide information about how fragile a product is, e.g. in the case of glassware; a degree of grip, which can provide information about how a product can be picked up, whether a product has a structure for picking up, the general grip of a product and / or whether a product tends to be easy to grip or rather slippery; and / or a degree of handling reliability, which can provide information about the success of a pick-and-place process, for example whether the product falls out of the product allocation device or not.

[0070] A "control signal" can be understood as a signal, e.g., an electrical, electronic, and / or digital signal, that can initiate, control, regulate, monitor, and / or influence the operation and / or behavior of a system, device, appliance, machine, and / or facility, and / or can signal something to it. It can serve as a means of communication, for example, for transmitting instructions, commands, instructions, and / or information, e.g., to initiate, execute, or perform certain actions, movements, actions, functions, and / or procedures. A control signal can comprise one or more pieces of information, for example, at least one property of at least one good, determined in particular based on the individualized representation. The control signal can also comprise one or more pieces of information regarding: how the good is to be picked up, in particular how it is to be grasped, e.g.,depending on an orientation, dimension, location, and / or position in the box, e.g., in relation to other goods; how the picked-up goods are to be inserted into the packaging, e.g., with respect to an orientation, dimension, location, and / or position, etc., and / or taking into account goods already inserted and / or goods yet to be inserted. Further information that the control signal may contain is derived from the following description.

[0071] According to the invention, the step of determining the at least one property associated with the individual product can comprise processing the generated individualized representation by means of the at least one AI system, wherein, in particular, the individualized representation is provided to the at least one AI system as input. Based on the individualized representation, the at least one AI system can make decisions, in particular regarding how the product is to be picked up from the chaotic box and / or how it is to be further processed, in particular regarding how the product is to be packaged, e.g., whether the picked-up product needs to be rotated in order to be able to package it (see also above).

[0072] According to the invention, the determination, in particular identification of the at least one property, can comprise an estimation and / or calculation, in particular based on the image data and / or the individualized representation, in particular based on one or more pixels and / or one or more images therein, in particular by means of the at least one AI system. The calculations or estimates that can be carried out within the scope of determining the properties of the goods can be based on mathematical methods, statistical analyses and / or other algorithmic approaches. The at least one AI system can, for example, due to its trained model and its ability to capture complex relationships, carry out precise calculations and / or estimates in order to obtain a statement about the goods, for example before the goods are picked up. In this way, the at least one AI system can, for example,a weight of the goods that are to be picked up and / or were identified in the individualized representation can be estimated. This process can be provided to the at least one AI system again, in particular continuously, e.g. for each item, as input, e.g. for training purposes. For example, goods with, e.g., a known weight can be passed to the at least one AI system, for which it carries out a calculation or estimate. In this case, the training data can correspond to the weight of the goods and the at least one AI system can be given feedback as to whether the weight determined, e.g., from the individualized representation, also corresponds to the actual weight, or how large the deviation is. On this basis, the at least one AI system can learn and, for example, improve its accuracy. The actual weight can also be determined via at least one sensor, whereby the at least one AI system, e.g.can also receive feedback automatically, e.g., without requiring user interaction. The same considerations apply analogously to other product characteristics mentioned herein.

[0073] According to the invention, the method may comprise:

[0074] - Obtaining one or more reference values ​​for the, in particular predefined, packaging and / or for the corresponding goods intake, in particular goods entry opening, in particular from a database, in particular from a cloud system;

[0075] - Compare whether at least one characteristic assigned to the individual product corresponds to the one or more reference values.

[0076] The one or more reference values ​​may include one or more of: a dimension, shape, material type and / or property, position, strength, stability, surface finish, assigned identifier, and / or weight. The assigned identifier may, for example, be an ID that can indicate whether the order is for an individual or a bulk order.

[0077] According to the invention, the method may comprise:

[0078] - Determine, on the basis of the comparison, i) whether the individual goods are suitable for the, in particular predefined, packaging and / or for the corresponding goods receptacle, in particular the goods entry opening, and / or ii) how the individual goods are to be inserted into and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening.

[0079] "Suitable" can be understood as meaning that the product and the packaging, especially predefined packaging, are coordinated. For example, this can mean that the individual product fits into a predefined packaging, e.g., with regard to its dimensions, e.g., length / width / depth. For example, a USB stick (product) can be suitable for a predefined packaging, e.g., with regard to its size, e.g., it fits, whereas a broom handle may not be suitable for the predefined packaging, e.g., a parcel, due to its length, e.g., because it is too long.

[0080] Determining "how" the individual goods are to be inserted can be done by taking into account a) a position, location, orientation, alignment, e.g., of the goods in the chaotic box, e.g., in relation to other goods, which may, for example, only allow limited possibilities for the individual goods to be accommodated by the goods allocation device; and / or b) a surface quality, dimension, shape, weight, stability, strength, hardness, and / or a material of the individual goods; and / or c) a final packaging and / or a packaging raw material, in particular with a specific and / or predetermined dimension, e.g., length, width, height, depth, thickness, in particular (packaging) wall thickness, a specific material, a specific shape, a specific weight, a specific surface quality, and / or a specific robustness, strength, and / or stability. Specific and / or predetermined customer-specific information, e.g.Whether the package is a single order or a bulk order, e.g., multiple items in one package, can also be included. Furthermore, how the individual item was picked up can be taken into account, e.g., taking into account the position, location, orientation, and alignment of the picked item on the item allocation device, which can, for example, determine a loading direction. For example, a picked ruler can be rotated before loading so that it can be loaded lengthwise, for example, instead of across.

[0081] According to the invention, the determination of steps i) and / or ii) may further include taking into account

[0082] - goods already placed in and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening, and / or

[0083] - nor into the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening, and / or the goods to be placed therein.

[0084] For example, the at least one KL system can, e.g. based on the individualized representation of the plurality of goods in a chaotic box, determine which goods are already arranged in a package, e.g. a number of goods, e.g. one, two, three, etc. and / or a type of goods, e.g. a razor and two contact lenses, etc.; and / or how these are arranged there, e.g. with regard to an orientation, position, location and / or direction of extension within the packaging. This can, for example, serve to optimally utilize a packaging size, e.g. for a group order. In each example mentioned herein, the at least one KL system can also take into account a gripping position of the goods, e.g. an orientation, e.g. after a goods allocation device has taken out or picked up, in particular gripped, an individual item from the, in particular chaotic, box.The gripping position can therefore influence the direction of insertion into the packaging. A gripping position can also be fed back to the at least one AI system (feedback) and serve as a starting point for training. This can, for example, be used to teach the at least one AI system that certain goods should be turned, pivoted, moved, and / or handled more carefully, so that they do not fall out of the gripper, and / or that certain goods should be held, secured, and / or gripped more securely, for example, to further secure them.

[0085] According to the invention, the comparing, the determining of steps i) and / or ii) and / or the taking into account a processing

[0086] -- the image data,

[0087] -- the generated individualized representation;

[0088] -- the one or more reference values; and / or

[0089] -- a feedback signal, in particular from a packaging station and / or a cloud system, by means of the at least one Kl system.

[0090] The "processing" may include the provision

[0091] -- the image data,

[0092] -- the generated individualized representation;

[0093] -- the one or more reference values; and / or

[0094] -- the feedback signal, as input for at least one AI system, e.g. for training or learning purposes.

[0095] According to the invention, the control signal can be configured to control a goods allocation device, in particular a robot, in particular a pick-and-place robot, in particular for packaging the individual goods in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening. According to the invention, the control signal can be configured to cause the goods allocation device to pick up, in particular to grip, the individual goods from the plurality of goods, in particular individually.

[0096] According to the invention, the control signal can indicate how the goods allocation device, on the basis of the at least one property, should pick up, in particular grip, the individual goods from the plurality of goods.

[0097] The indication "how" the individual goods are to be picked up can be made taking into account a) a position, location, orientation, alignment, e.g. of the goods in the chaotic box, e.g. in relation to other goods, which may only allow limited picking up of the individual goods by means of the goods allocation device, for example if the goods are on top of each other, next to each other, underneath each other, overlapping and / or covering each other; and / or b) a surface quality, dimension, shape, weight, stability, strength, hardness and / or material of the individual goods. For example, it may be necessary to

[0098] It is advisable to grasp goods at a stronger or more stable point, e.g., a thicker point, rather than at a fragile or easily breakable / damaged point, e.g., made of glass. It can also be taken into account that goods are more likely to break in some places. At least one AI system can make such decisions, e.g., based on the individualized display and / or image data and / or previous training data.

[0099] The control signal can indicate how the goods allocation device handles the picked goods

[0100] - into the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening; and / or to be arranged therein, in particular with regard to the at least one property assigned to the individual goods.

[0101] The "how" of introducing the individual goods has already been described above; this also applies here.

[0102] The method according to the invention may further comprise:

[0103] Receiving a feedback signal, in particular from the packaging station and / or the cloud system, in particular after the individual goods have been introduced and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening.

[0104] A cloud system can be understood as an infrastructure that can comprise one or more (e.g., virtual or virtualized) components or computing resources, such as a server, a storage device, a database, and / or a network for communication between the components. The cloud system can be centralized or decentralized.

[0105] According to the invention, the feedback signal can indicate whether the packaging of the individual goods in the, in particular predefined, packaging and / or the goods receptacle corresponding thereto and / or formed therein, in particular the goods input opening, was successful and / or how the individual goods were arranged therein, in particular with regard to the at least one property assigned to the individual goods.

[0106] "Successful" can be understood as meaning that the goods are suitable for the packaging, e.g., the final packaging and / or the packaging raw material, e.g., with regard to a dimension, and in particular that they can be packaged in / with it. This can also indicate that the goods have been packaged in the packaging.

[0107] The indication "how" the individual goods were arranged therein can be determined taking into account a1) a position, location, orientation, alignment, e.g. of the goods in the chaotic box, e.g. in relation to other goods, which may only allow limited possibilities for picking up the individual goods by means of the goods allocation device; and / or a2) a position, location, orientation, alignment, e.g. of the picked up goods on the goods allocation device, which maycan determine an insertion direction and may require a change in position, in particular a rotation; and / or b) a surface quality, dimension, shape, weight, stability, strength, hardness and / or a material of the individual and / or picked up goods and / or c) a final packaging and / or a packaging raw material with a specific and / or predetermined dimension, for example length, width, height, depth, thickness, in particular (packaging) wall thickness, a specific material, a specific shape, a specific weight, a specific surface quality and / or a specific robustness, strength and / or stability. Specific and / or predetermined customer-specific information, e.g. whether the package is an individual order or a collective order, e.g. several goods in one package, can also be included.According to the invention, the feedback signal can be provided to the at least one AI system as input, in particular before and / or after generating the personalized representation of the plurality of goods. For example, this can be done automatically and / or manually for training purposes, e.g., through supervised and / or unsupervised training. Other training data can also be used.

[0108] The invention relates to a computer-implemented method, in particular according to one or more of the aspects and / or embodiments described herein, for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations, comprising:

[0109] - generating, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualised representation of the plurality of goods,

[0110] - Determine, in particular identify, on the basis of the individualized representation,

[0111] -- an individual product from the plurality of products, and / or

[0112] -- of at least one characteristic associated with the individual product; and,

[0113] - Outputting a control signal based on the at least one specific property.

[0114] A "packing station" can represent an area, location, and / or zone where specific devices, equipment, machines, and / or facilities for packing goods are present, e.g., requiring a specific packaging process, a specific packaging material, and / or a specific packaging type. A packing station can, for example, be an area where an employee performs manual packing, e.g., because at least one AI system has determined and / or decided that the goods are easily breakable, e.g., fragile, and should be packed manually, e.g., by an employee. Another example is a parcel station, where goods are packed, e.g., in a folded box, or a station for letters of specific dimensions, e.g., depending on the postage, especially maxi letters or large letters, etc.

[0115] The control signal can indicate to which of the plurality of packaging stations the one or more goods should be transferred, for example, depending on the at least one property of the goods. The at least one property can include a position, location, orientation, alignment, surface finish, size, dimension, shape, weight, stability, strength, hardness, a degree of fragility, a degree of grip, a degree of handling reliability, and / or a material of the individual goods.

[0116] The step of generating the individualized representation of the plurality of goods may comprise processing the image data by means of at least one artificial intelligence (Kl) system, wherein in particular the image data are provided to the at least one Kl system as input and wherein in particular the image data are obtained from at least one image recording device.

[0117] The step of determining the at least one property associated with the individual product may comprise processing the generated individualized representation by means of the at least one AI system, wherein in particular the individualized representation is provided to the at least one AI system as input.

[0118] Determining, in particular identifying, the at least one property may comprise an estimation and / or calculation, in particular based on the image data and / or the individualized representation, in particular based on one or more pixels and / or one or more images therein, in particular by means of the at least one Kl system.

[0119] The method according to the invention may comprise:

[0120] - Obtaining one or more reference values ​​for at least two of the plurality of packaging stations, in particular from a database, in particular from a cloud system;

[0121] - Compare whether at least one characteristic assigned to the individual product corresponds to the one or more reference values.

[0122] The plurality of packaging stations may comprise at least two packaging stations that differ in terms of a packaging process, packaging material, and / or packaging type. However, identical / identical packaging stations are also possible.

[0123] The one or more reference values ​​for at least two of the plurality of packaging stations may comprise one or more of:

[0124] - a dimension of a goods receptacle, in particular a goods input opening of at least two packaging stations, and / or a shape, material type and / or property, strength, stability, surface quality of a packaging material of the at least two packaging stations, and / or

[0125] - a dimension and / or shape of a final packaging and / or packaging raw material formed by the at least two packaging stations.

[0126] The method according to the invention may comprise:

[0127] - Determining, on the basis of the comparison, i) for which specific one of the plurality of packaging stations the individual product is suitable, and / or ii) how the individual product is to be handed over to the specific packaging station, in particular on the basis of the at least one property.

[0128] "Suitable" can be understood to mean that the product and the product receptacle, in particular the product entry opening of the packaging station, are coordinated for packaging. For example, that the individual product, e.g., fits into a product receptacle, in particular a product entry opening, e.g., a product entry slot, e.g., with a predetermined dimension, in terms of its dimensions, e.g., size. For example, a USB stick (product) may be suitable for a product receptacle, in particular a product entry opening, in terms of its size, e.g., fitting, whereas a vase may not be suitable for the product receptacle, in particular the product entry opening, e.g., due to its shape / width, e.g., because it is too wide / thick.

[0129] The indication "how" the individual goods are to be handed over to the specific packaging station can be determined taking into account a1) a position, location, orientation, alignment, e.g. of the goods in the chaotic box, e.g. in relation to other goods, which may only allow limited picking up of the individual goods by means of the goods allocation device; and / or a2) a position, location, orientation, alignment, e.g. of the picked up goods at the goods allocation device, which maycan specify an insertion direction; and / or b) a surface quality, dimension, shape, weight, stability, strength, hardness and / or a material of the individual goods; and / or the picked up goods and / or c) the goods receptacle, in particular the goods input opening, with a specific and / or predetermined dimension, for example length, width, height, depth, thickness, in particular (packaging) wall thickness, a specific material, a specific shape, a specific weight, a specific surface quality and / or a specific robustness, strength and / or stability. Specific and / or predetermined customer-specific information, e.g. whether the package is an individual order or a collective order, e.g. several goods in one package, can also be included.

[0130] According to the invention, the comparing and / or determining of steps i) and / or ii) may involve processing

[0131] -- the image data,

[0132] -- the generated individualized representation;

[0133] -- the one or more reference values; and / or

[0134] -- a feedback signal, in particular from the one or more packaging stations and / or a cloud system, by means of the at least one Kl system.

[0135] The control signal can be configured to control a goods allocation device, in particular a robot, in particular a pick-and-place robot, in particular for picking up, in particular gripping, the individual item of goods from the plurality of goods and / or for transferring the individual item of goods to the specific one of the plurality of packaging stations, in particular based on the at least one property. The control signal can also be configured to cause the goods allocation device to introduce the goods into the goods receptacle, in particular the goods input opening, of a packaging station, e.g., to insert it into a goods input slot and / or, if necessary, to bring the picked up goods into the corresponding position, e.g., to align and / or rotate them accordingly.

[0136] A goods allocation device can be a robot, in particular a pick-and-place robot, for packaging one or more goods in a, in particular predefined, packaging and / or in a corresponding packaging and / or for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations, in particular for introducing and / or arranging in a corresponding goods receptacle, in particular a goods input opening, of the specific one of the plurality of packaging stations.

[0137] A system according to the invention for packaging one or more goods can comprise the goods allocation device according to the invention, which implements one or more of the inventive methods described herein, and at least one image recording device, in particular with at least one camera, for obtaining the image data. The system can further comprise the at least one KI system described herein.

[0138] The system may further comprise at least one packaging station, or at least two packaging stations that differ with respect to a packaging process, a packaging material and / or a packaging type.

[0139] The system may further comprise at least one sensor device which is configured

[0140] -- to determine at least one further characteristic associated with the individual goods, and

[0141] -- to provide the at least one further property to the at least one Kl system as, in particular, additional, input.

[0142] The at least one sensor device can be located on the goods allocation device, in particular be attached to a goods receiving end point. For example, the at least one AI system, as already described, can carry out an estimate and / or calculation before and / or during the picking of the goods from the chaotic box, e.g. in order to estimate at least one property of the individual goods. The sensor device can, for example, provide additional information, e.g. after the picking of the goods, in order to verify the estimated and / or calculated data or information, in particular by means of the at least one AI system. For example, the at least one sensor device can comprise one or more of: at least one camera, at least one weight measuring device, in particular a scale, and / or at least one laser measuring device.For example, after picking up the goods from the chaotic box, a weight measuring device can perform a weight measurement to confirm the pre-estimation and / or pre-calculation of the at least one AI system, or provide the measured result to the at least one AI system as input and / or feedback. The estimation and / or calculation of the at least one property of the individual goods can be confirmed (or not) in order to (further) train the at least one AI system in this regard and improve its accuracy.

[0143] The system according to the invention can further comprise a cloud system that is communicatively coupled to the at least one K1 system, wherein the cloud system comprises a database. In the database, for example, customer-related data can be stored that is relevant, for example, for the delivery of the package, e.g., name, address and / or contents, e.g., an individual order or a collective order. Both the one or more packaging stations and the goods allocation device can each be connected to the cloud system, for example, all components of the system according to the invention, in order to enable communication. For example, a communication connection, e.g.,for exchanging information, one or more of: Ethernet, Internet, LAN (Local Area Network), WLAN (Wireless Local Area Network), VPN (Virtual Private Network), wireless technologies such as Bluetooth, Wi-Fi Direct, Near Field Communication (NFC), the Internet and / or mobile networks (e.g. GSM, 3G, 4G, 5G etc.).

[0144] BRIEF DESCRIPTION OF THE DRAWINGS

[0145] Embodiments of the invention will now be described with reference to the accompanying drawings. In order to understand the above-mentioned features of the present disclosure in detail, a more detailed description of the disclosure, briefly summarized above, can be obtained by reference to embodiments. The accompanying drawings relate to embodiments of the disclosure and are described below:

[0146] Fig. 1A shows a flowchart of a method according to the invention according to an embodiment;

[0147] Fig. 1B shows the steps of generating the individualized representation of the method according to the invention according to the embodiment of Fig. 1A;

[0148] Fig. 1C shows the steps of determining the individual product and / or the at least one property of the method according to the invention according to the embodiment of Fig. 1A;

[0149] Fig. 1D shows the output of the control signal of the method according to the invention according to the embodiment of Fig. 1A;

[0150] Fig. 1E shows the use of at least one or at least two Kl systems and output of the control signal of the method according to the invention according to the embodiment of Fig. 1A;

[0151] Fig. 2 shows a flowchart of a method according to the invention according to another embodiment; Fig. 3 shows image data obtained by means of an image recording device and / or an individualized representation of a plurality of goods obtained by means of the method according to the invention according to one or more embodiments;

[0152] Fig. 4 shows a system according to the invention for packaging one or more goods with a goods allocation device, in particular a robot, in particular a pick-and-place robot, and a plurality of packaging stations according to one or more embodiments.

[0153] DESCRIPTION OF PREFERRED EMBODIMENTS

[0154] In the following, the invention will be explained in more detail with reference to embodiments shown in the drawings, wherein in all drawings, essentially functionally identical elements have the same reference numerals.

[0155] The drawings are schematic drawings and are not to scale. Some elements in the drawings may have exaggerated dimensions to emphasize aspects of the present disclosure and / or for presentation clarity. For simplicity, identical reference numerals are used to identify identical elements commonly included in the drawings. It is contemplated that elements and features of one embodiment may be advantageously incorporated into other embodiments without further notice. In general, only the differences with respect to individual embodiments will be described.

[0156] Each embodiment is provided to illustrate the disclosure and is not intended to limit the disclosure. Furthermore, features shown or described as part of one embodiment may be used in conjunction with other embodiments.

[0157] Embodiments may be used to create another embodiment. The description is intended to encompass such modifications and variations.

[0158] Fig. 1A shows a flowchart of a method 100 according to the invention according to an exemplary embodiment for packaging one or more goods in a, in particular predefined, packaging and / or in a corresponding goods receptacle, in particular a goods input opening. The method 100 comprises generating 110, based on image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualized representation of the plurality of goods; determining, in particular identifying, based on the individualized representation, an individual good 120 from the plurality of goods and / or at least one property 130 assigned to the individual good; and outputting 140 a control signal based on the at least one determined property.

[0159] Optionally, the method may also comprise controlling a goods allocation device for packaging the goods in a, in particular predefined, packaging and / or a packaging corresponding thereto.

[0160] Fig. 1B shows the steps of generating 110 the individualized representation of the method 100 according to the invention according to the embodiment from Fig. 1A. For example, the image data 150 can be obtained using an image recording device (see Fig. 3), for example at least one camera. The image data 150 can comprise a two-dimensional representation (2D) and / or a three-dimensional representation (3D), which can comprise, for example, one or more pixels and / or one or more images of the, in particular chaotic, box, in particular its contents, in particular the plurality of goods. The image data 150 can be provided to the AI ​​system 170 as input. In addition, training data 160 can optionally also be provided, e.g., to train the AI ​​system 170. The training data 160 can show a plurality of goods, e.g., bulk goods, which are arranged, in particular chaotically and / or randomly, in a box.

[0161] In addition, the Kl system 170 can be coupled, in particular communicatively, to a cloud system 197, for example via a communication connection, e.g., for exchanging information, such as Ethernet, Internet, LAN (Local Area Network), WLAN (Wireless Local Area Network), VPN (Virtual Private Network), wireless technologies such as Bluetooth, Wi-Fi Direct, Near Field Communication (NFC), the Internet, and / or mobile networks (e.g., GSM, 3G, 4G, 5G, etc.). The cloud system can include a database.

[0162] The cloud system 197 can provide the AI ​​system 170 with customer-specific information (e.g., name, address, whether it is a single or bulk order, etc.) and / or training data, in particular additional training data, e.g., image data or other data. Furthermore, reference values ​​can be provided, whereby feedback to the AI ​​system via the cloud system 197 is also conceivable, e.g., based on the generated personalized representation.

[0163] According to the invention, the AI ​​system 170 can generate an individualized representation 180 of the plurality of goods from the image data 150, optionally the training data 160, optionally the data from the cloud system 197. The individualized representation 180 can be understood as a (2D and / or 3D) representation, e.g., of the plurality of goods, e.g., within the chaotic box, in which one, e.g., first, item of goods is recognizable, identifiable, and / or detectable separately, individualized, segmented, rasterized, isolated, and / or detached from a, e.g., second, in particular identical or different, item of goods (see, e.g., Fig. 3). The individualized representation 180 can, for example, comprise a three-dimensional (3D) representation of the plurality of goods, e.g., including a spatial depth that comprises, for example, one or more pixels and / or one or more images, e.g., of the chaotic box, in particular its contents, in particular of the plurality of goods.For example, the goods can be displayed individually and separately from one another, including with regard to depth, e.g., when the goods are stacked on top of one another. Generating the individualized display can include processing the image data using the AI ​​system 170. It is also possible to subsequently transfer the individualized display to the AI ​​system 170 as training data, for example, to improve the generation of the individualized display, in particular to improve recognition of the individual goods within the box.

[0164] Fig. 1C shows the steps of determining 120, 130 the individual product and / or the at least one property of the inventive method 100 according to the embodiment from Fig. 1A. As shown in Fig. 1C, the individualized representation 180 can be provided to the AI ​​system 170 as input, for example the individual representation 180 generated in Fig. 1B, in order to determine, in particular to identify, an individual product from the plurality of products and / or to determine, in particular to identify, at least one property that is assigned to the individual product. This can include processing the generated individualized representation 180 by means of the AI ​​system 170, wherein in particular the individualized representation 180 is provided to the AI ​​system 170 as input. It should be understood that for the determination 120 of the individual product and / or the at least one property as shown in Fig. 1C, the image data 150 from Fig.1 B can be used, e.g., additionally, to provide the AI ​​system 170 with more data / information. Based on the individualized representation 180, the AI ​​system 170 can make decisions, in particular how the goods are to be picked up from the chaotic box, e.g., at which point the goods are to be picked up or in relation to other goods, and / or how to proceed with them, in particular how the goods are to be packaged, e.g., whether the picked up goods need to be turned in order to be packaged or inserted into the packaging.

[0165] The determination 120, 130, in particular identifying the at least one property, can comprise an estimation and / or calculation, in particular based on the image data and / or the individualized representation 180, in particular based on one or more pixels and / or one or more images therein, in particular by means of the AI ​​system 170. The calculations or estimates that can be carried out as part of the property determination of the goods can be based on mathematical methods, statistical analyses and / or other algorithmic approaches. The AI ​​system 170 can, for example, due to its trained model and its ability to capture complex relationships, carry out precise calculations and / or estimates in order to obtain a statement about the goods, for example before the goods are picked up. In this way, the AI ​​system 170 can, for example,The weight of the item to be picked up and / or identified in the individualized display can be estimated. This process can be provided to the AI ​​system 170 again, in particular continuously, e.g., for each item, as input, e.g., for training purposes.

[0166] Fig. 1D shows the output of the control signal 140 of the inventive method 100 according to the embodiment of Fig. 1A. To output the control signal, the AI ​​system 170 can use, in particular process, one or more of: the image data 150, training data 160, the individualized representation 180, one or more reference values ​​190, e.g., information from the cloud system 197, a feedback signal 195.

[0167] The control signal can be configured to cause a goods allocation device (see Fig. 4; 405) to pick up, in particular grip, the individual item from the plurality of items, in particular individually. The control signal can indicate how the goods allocation device, based on the at least one property, should pick up, in particular grip, the individual item from the plurality of items. The control signal can indicate how the goods allocation device should insert and / or arrange the picked-up item into the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening, in particular with respect to the at least one property assigned to the individual item.

[0168] The control signal can comprise one or more pieces of information on: how the goods are to be picked up, in particular gripped, e.g. depending on an orientation, dimension, location and / or position in the box, e.g. in relation to other goods; how the picked up goods are to be inserted into the packaging, e.g. with regard to an orientation, dimension, location and / or position, etc. and / or taking into account goods already inserted and / or still to be inserted.

[0169] The Cl system 170 can then compare whether the at least one property assigned to the individual product corresponds to the one or more reference values ​​190, e.g. to determine i) whether the individual product is suitable for the, in particular predefined, packaging and / or for the corresponding goods receptacle, in particular the goods input opening, and / or ii) how the individual product is to be introduced into the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening, and / or arranged therein.

[0170] In this context, goods already introduced into and / or arranged in the, in particular predefined, packaging and / or the goods receptacle corresponding thereto, in particular the goods input opening, and / or goods still to be introduced into and / or arranged in the, in particular predefined, packaging and / or the goods receptacle corresponding thereto, in particular the goods input opening, can also be taken into account by the Kl system 170.

[0171] For the sake of clarity, reference is made to the above description for details on "how" the goods are to be picked up, inserted and / or arranged and / or whether the goods are "suitable".

[0172] Fig. 1E shows the use of at least one Kl system 170 or at least two Kl systems, i.e. a first Kl system 170A and a second Kl system 170B, which can be communicatively connected to one another, for example, as indicated by the arrow 170C, as well as the output of the control signal 140 of the method according to the invention according to the exemplary embodiment from Fig. 1A. It should be noted that this embodiment can be combined with any embodiment described herein, e.g. an embodiment in which, for example, only one Kl system is disclosed and / or described, this one Kl system can be replaced or exchanged by the at least two Kl systems, i.e. the first Kl system 170A and the second Kl system 170B.

[0173] As can be seen, a first AI system 170A can receive the image data 150 and / or the individualized representation 180, e.g., as input or first input. After processing, the first AI system can generate a first output, which can either be stored or provided to the second AI system 170B as input or second input. The second AI system 170B can, in particular based on the second input, generate a second output, which can correspond to the control signal 140 for controlling the goods allocation device and / or can be integrated into the control signal 140.

[0174] Fig. 2 shows a flowchart of a method 200 according to the invention according to a further embodiment, which can be combined with any aspect and / or embodiment described herein, for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations, comprising:

[0175] - generating 210, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualized representation of the plurality of goods,

[0176] - Determine, in particular identify, on the basis of the individualized representation,

[0177] -- an individual product 220 from the plurality of products, and / or

[0178] -- of at least one characteristic 230 associated with the individual product; and,

[0179] - Outputting 240 a control signal based on the at least one specific property.

[0180] Optionally, the method may also comprise controlling a goods allocation device for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations.

[0181] It should be noted that method 200 differs from method 100, particularly in the type of control signal. While the control signal generated by method 100 serves to control a goods allocation device for packaging the goods in a package and / or a corresponding goods receptacle, in particular the goods input opening, of a packaging station, the control signal 200 generated by method 200 can serve to control a goods allocation device for allocating and / or packaging one or more goods using a specific one of a plurality of packaging stations (see Fig. 4).

[0182] The control signal generated by method 200, in particular by means of the KI system 170, can be configured to control a goods allocation device, in particular a robot, in particular a pick-and-place robot, in particular for picking up, in particular gripping, the individual item of goods from the plurality of goods and / or for transferring the individual item of goods to the specific one of the plurality of packaging stations, in particular based on the at least one property. The control signal can also be configured to cause the goods allocation device to insert the goods into the goods receptacle, in particular the goods input opening, of a packaging station, e.g., to insert it into a goods input slot and / or, if necessary, to bring the picked up goods into the corresponding position, e.g., to align and / or rotate them accordingly.

[0183] Fig. 3 shows image data 150 obtained by means of an image recording device 310, in particular at least one camera 311, and / or an individualized representation 180 of a plurality of goods obtained by means of the method 100, 200 according to the invention, in particular by means of the KI system 170.

[0184] For example, the camera 311 can capture, record, and / or capture 312 the contents of a, in particular, chaotic box 320. The image data 150 shows the contents of the chaotic box, in particular a plurality of goods randomly arranged therein. As can be seen, the goods have different properties; however, they can also have identical properties. For example, the box 320 contains scissors 320A, a pen 320B, and a pack of sticky tape rolls 320C, which are referred to here as examples. In addition, there are also further / different goods in the box 320. The majority of the goods in the box 320 are located on top of one another, next to one another, below one another, overlap one another, cover one another, are located on the back, and / or have a, in particular random, alignment, position, location, and / or orientation in the container 320 or the box 320.For example, additional goods are located below the scissors 320A, the pen 320B, and the pack of adhesive tape rolls 320C. The pack of adhesive tape rolls 320C is also arranged with the back facing up compared to the scissors 320A and the pen 320B.

[0185] As can be seen in Fig. 3, the AI ​​system 170 can generate an individualized representation 180 based on the image data 150. The individualized representation 180 shows the contents of the box 320, in which a first product, e.g. the scissors 320A, is shown separately, individualized, segmented, rasterized, isolated and / or detached from a second product, e.g. the pen 320B. The products 320A, 320B and 320C are, for example, shown individually. In the specific example in Fig. 3, the individualized representation 180 was generated using bounding boxes 320A', 320B' and 320C. However, other ways of generating the individualized representation are also conceivable.

[0186] As can be further seen in Fig. 3, in particular the KL system 170 can determine at least one property associated with the individual product 320A, 320B and / or 320C based on the individualized representation 180. In particular, a position, location, orientation, alignment, surface quality, dimension, shape, weight, stability, strength, hardness, a degree of fragility, a degree of grip, a degree of handling reliability and / or a material of the individual product 320A, 320B and / or 320C can be determined. A particular article, e.g., a pair of scissors 320A, can be recognized and distinguished from the pen 320B and / or the pack of adhesive tape rolls 320C.

[0187] Fig. 4 shows a system 400 according to the invention for packaging one or more goods 402, 404 with a goods allocation device 405, in particular a robot 405, in particular a pick-and-place robot 405, and a plurality of packaging stations 420-450. The system can be used to carry out or implement the aspects and / or embodiments of the invention, in particular the methods of the invention, in particular individually and / or in combination.

[0188] The inventive system 400 for packaging one or more goods 402, 404 can comprise the inventive goods allocation device 405, which implements one or more of the inventive methods described herein, and at least one image recording device 310 (see Fig. 3), in particular with at least one camera, for obtaining the image data. The image recording device can, for example, be arranged directly on the goods allocation device 405 or at another location where the at least one camera is capable of, for example, capturing, recording, and / or recording the contents of a particularly chaotic box 320. The system 400 can further comprise the AI ​​system 170 described herein. As can be seen in Fig. 4, the goods allocation device 405 can be equipped with or comprise an AI system 170. However, the AI ​​system 170 can also be placed at another location.The system 400 further comprises at least two packaging stations 420-450 that differ in terms of a packaging process, a packaging material, and / or a packaging type. However, identical / identical packaging stations are also possible.

[0189] The plurality of goods 402, 404 can be located, for example, in a container, a box and / or a carton, in particular in a chaotic box and / or on a conveyor belt 410, which can transport the goods 402, 404 and / or the container, the box and / or the carton, in particular in the chaotic box, to the goods allocation device 405 and / or can make them available to the latter. This, in particular chaotic, box can therefore be located, for example, on the conveyor belt 410, or at least at a location at which the goods allocation device 405 can access the box, in particular its contents, in particular the plurality of goods 402, 404, e.g., in order to individually remove goods therefrom. It is also conceivable for the box to be located on a shelf and / or to be provided and / or transported by a means of transport, for example, a trolley.

[0190] The system may further comprise at least one sensor device (not shown) configured to determine at least one further property associated with the individual item of goods 402, 404, and to provide the at least one further property to the AI ​​system 170 as, in particular, additional input. The at least one sensor device may be located on the goods allocation device 405, in particular attached to a goods receiving end point, e.g., to a gripper, in particular a goods allocation device, for example, a pick-and-place robot. For example, the AI ​​system 170, as already described, may perform an estimation and / or calculation before and / or during the picking of the goods 402, 404 from the chaotic box, e.g., to estimate or pre-estimate at least one property of the individual item of goods 402, 404. The sensor device may, for example, provide additional information, e.g.,after picking up the goods 402, 404, in order to verify the estimated and / or calculated data or information, in particular by means of the AI ​​system 170. For example, the at least one sensor device can comprise one or more of: at least one camera, at least one weight measuring device, in particular a scale, and / or at least one laser measuring device. For example, after picking up the goods 402, 404, a weight measuring device can perform a weight measurement in order to confirm the pre-estimation and / or pre-calculation of the AI ​​system 170, or provide the measured result to the AI ​​system as input and / or feedback.

[0191] In Fig. 4, a plurality of packaging stations 420-450 are shown, which have a specific packaging process, a specific packaging material and / or a specific packaging type. For example, the packaging station 420 can be an area where an employee performs manual packaging, e.g. because the AI ​​system 170 has detected, determined and / or decided that the goods 402, 404 are easily breakable, e.g. fragile, and should be packaged manually, e.g. by an employee. The goods allocation device 405 can then transfer the goods 402, 404 to the packaging station 420. From this packaging station, for example, manual feedback can be fed back to the AI ​​system, although automatic feedback is also possible. Another example is a parcel station 430, where the goods, e.g.in a final packaging, such as a folded carton, envelope, or the like, or a station for letters of specific dimensions, e.g., depending on the postage, especially Maxi Letters 440 or Large Letters 450, e.g., with other dimensions. From this packaging station, for example, automatic feedback can be sent back to the KL system, although manual feedback is also possible.

[0192] The control signal can indicate to which of the plurality of packaging stations 420-450 the one or more goods 402, 404 should be transferred, for example, depending on the at least one property of the goods. The at least one property can include a position, location, orientation, alignment, surface finish, size, dimension, shape, weight, stability, strength, hardness, a degree of fragility, a degree of grip, a degree of handling reliability, and / or a material of the individual goods 402, 404.

[0193] For example, this can be done by comparing whether the at least one property associated with the individual product matches the one or more reference values. The one or more reference values ​​can be obtained from the cloud system and include one or more of:

[0194] - a dimension of a goods receptacle, in particular a goods input opening, of the at least two packaging stations 420-450, and / or a shape, material type and / or property, strength, stability, surface quality of a packaging material of the at least two packaging stations 420-450, and / or - a dimension and / or shape of a final packaging and / or a packaging raw material which is formed by the at least two packaging stations 420-450.

[0195] As already mentioned, the one or more packaging stations 420-450, the AI ​​system 170 and / or the goods allocation device 405 can be connected to the cloud system 197, in particular communicatively, e.g., via a communication connection (for the sake of clarity, only one connection is shown graphically in Fig. 4). In this way, the one or more packaging stations 420-450 can provide the AI ​​system with feedback and / or additional input, in particular whether the packaging of the individual goods by means of the specific one of the plurality of packaging stations 420-450 was successful and / or how the individual goods were arranged therein, in particular with regard to the at least one property assigned to the individual goods. This allows the AI ​​system 170 to be improved, in particular its accuracy, reliability and / or speed.The feedback can be provided automatically, for example, via a communication connection, or, as in the case of the packaging station 420, manually, e.g., by the employee. This can be used, in particular, to further train and / or improve the AI ​​system 170.

[0196] While the foregoing is directed to embodiments of the disclosure, other and further embodiments of the disclosure may be devised without departing from the basic scope thereof, the scope being determined by the following claims.

[0197] List of reference symbols

[0198] 100 procedures

[0199] 110 Creating individualized representation

[0200] 120 Determining individual goods

[0201] 130 Determining at least one property

[0202] 140 Output control signal

[0203] 150 image data

[0204] 160 training data

[0205] 170 Kl system

[0206] 170A First Class System

[0207] 170B Second air system

[0208] 170C connection

[0209] 180 Individualized display

[0210] 190 reference values

[0211] 195 Feedback signal

[0212] 197 Cloud system

[0213] 200 procedures

[0214] 210 Creating individualized representation

[0215] 220 Determining individual goods

[0216] 230 Determining at least one property

[0217] 240 Output control signal

[0218] 310 Image recording device

[0219] 311 Camera

[0220] 312 Capture

[0221] 320 boxes

[0222] 320A scissors

[0223] 320B pen

[0224] 320C Tesa tape roll

[0225] 320A'-320C Bounding Box

[0226] 400 system

[0227] 402, 404 goods

[0228] 405 Goods allocation device

[0229] 410 boxes / conveyor belt

[0230] 420-450 packaging stations

Claims

PATENT CLAIMS 1. Computer-implemented method for packaging one or more goods in a, in particular predefined, packaging and / or in a corresponding goods receptacle, in particular a goods input opening, a packaging station, comprising: - generating, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualised representation of the plurality of goods, - Determine, in particular identify, on the basis of the individualized representation, -- an individual product from the plurality of products, and / or -- of at least one characteristic associated with the individual product; and, - Outputting a control signal based on the at least one specific property.

2. Method according to claim 1, wherein the particularly predefined packaging - a packaging raw material, in particular a substantially flat one; and / or - a final packaging formed, in particular folded, from the packaging raw material, in particular a shipping bag, preferably an envelope or a package.

3. The method according to any one of claims 1-2, wherein the at least one property comprises a position, location, orientation, alignment, surface finish, dimension, shape, weight, stability, strength, hardness, a degree of fragility, a degree of grip, a degree of handling reliability and / or a material of the individual product.

4. The method according to any one of claims 1-3, wherein the step of generating the individualized representation of the plurality of goods comprises a Processing the image data by means of an artificial intelligence, AI, system, wherein in particular the image data are provided to the AI system as input and wherein in particular the image data are obtained from an image recording device.

5. The method according to any one of claims 1-4, wherein the step of determining the at least one property associated with the individual product comprises a Processing the generated individualized representation by means of the Kl system, wherein in particular the individualized representation is provided to the Kl system as input.

6. Method according to one of claims 1-5, wherein the determining, in particular identifying, of the at least one property comprises an estimation and / or calculation, in particular on the basis of the image data and / or the individualized representation, in particular on the basis of one or more pixels and / or one or more images therein, in particular by means of the AI system.

7. The method according to any one of claims 1-6, further comprising: - Obtaining one or more reference values for the, in particular predefined, packaging and / or for the corresponding goods intake, in particular the goods input opening, of the packaging station, in particular from a database, in particular from a cloud system; - Compare whether at least one characteristic assigned to the individual product corresponds to the one or more reference values.

8. The method of claim 7, wherein the one or more reference values comprise one or more of: a dimension, shape, material type and / or property, location, position, alignment, orientation, strength, stability, surface finish, associated identifier and / or a weight.

9. The method according to any one of claims 7-8, further comprising: - Determine, on the basis of the comparison, i) whether the individual goods are suitable for the, in particular predefined, packaging and / or for the corresponding goods receptacle, in particular the goods entry opening, and / or ii) how the individual goods are to be inserted into and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening.

10. The method according to claim 9, wherein determining steps i) and / or ii) further comprises taking into account - goods already placed in and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening, and / or - nor the goods to be placed in and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening.

11. Method according to one of claims 9-10, wherein the comparing, the determining of steps i) and / or ii) and / or the taking into account a processing -- the image data, -- the generated individualized representation; -- the one or more reference values; and / or -- a feedback signal, in particular from a packaging station and / or a cloud system, by means of the Kl system.

12. Method according to one of claims 1-11, wherein the control signal is set up to control a goods allocation device, in particular a robot, in particular a pick-and-place robot, in particular for packaging the individual goods in the, in particular predefined, packaging and / or the goods receptacle corresponding thereto, in particular the goods input opening.

13. The method according to claim 12, wherein the control signal is configured to cause the goods allocation device to pick up, in particular to grip, the individual goods from the plurality of goods, in particular individually.

14. The method according to any one of claims 12-13, wherein the control signal indicates how the goods allocation device, based on the at least one property, should pick up, in particular grip, the individual goods from the plurality of goods.

15. Method according to one of claims 12-14, wherein the control signal indicates how the goods allocation device assigns the picked-up goods - into the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods entry opening; and / or to be arranged therein, in particular with regard to the at least one property assigned to the individual goods.

16. The method of claim 15, further comprising: Receiving a feedback signal, in particular from the packaging station and / or a cloud system, in particular after the individual goods have been introduced and / or arranged in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening.

17. The method according to claim 16, wherein the feedback signal indicates whether the packaging of the individual goods in the, in particular predefined, packaging and / or the corresponding goods receptacle, in particular the goods input opening, was successful and / or how the individual goods were arranged therein, in particular with regard to the at least one property assigned to the individual goods.

18. Method according to one of claims 16-17, wherein the feedback signal is provided to the AI system as input, in particular before, during and / or after generating the individualized representation of the plurality of goods.

19. A computer-implemented method, in particular according to one of claims 1-18, for allocating and / or packaging one or more goods by means of a specific one of a plurality of packaging stations, comprising: - generating, on the basis of image data representing a plurality of goods with at least two, in particular different or identical, properties, an individualised representation of the plurality of goods, - Determine, in particular identify, on the basis of the individualized representation, -- an individual product from the plurality of products, and / or -- of at least one characteristic associated with the individual product; and, - Outputting a control signal based on the at least one specific property.

20. The method according to claim 19, wherein the control signal indicates, in particular selects, to which of the plurality of packaging stations the one or more goods are to be transferred.

21. The method according to any one of claims 19-20, wherein the at least one property comprises a position, location, orientation, alignment, surface texture, size, Dimension, shape, weight, stability, strength, hardness, a degree of fragility, a degree of grip, a degree of handling reliability and / or a material of the individual product.

22. A method according to any one of claims 19-21, wherein the step of generating the individualized representation of the plurality of goods comprises a Processing the image data by means of an artificial intelligence (AI) system, wherein in particular the image data are provided to the AI system as input and wherein in particular the image data are obtained from an image recording device.

23. The method according to any one of claims 19-22, wherein the step of determining the at least one property associated with the individual product comprises processing the generated individualized representation by means of the AI system, wherein in particular the individualized representations are provided to the AI system as input.

24. Method according to one of claims 19-23, wherein the determining, in particular identifying, of the at least one property comprises an estimation and / or calculation, in particular on the basis of the image data and / or the individualized representation, in particular on the basis of one or more pixels and / or one or more images therein, in particular by means of the AI system.

25. The method according to any one of claims 19-24, further comprising: - Obtaining one or more reference values for at least two of the plurality of packaging stations, in particular from a database, in particular from a cloud system; - Compare whether at least one characteristic assigned to the individual product corresponds to the one or more reference values.

26. The method according to any one of claims 19-25, wherein the plurality of packaging stations comprise at least two packaging stations that differ with respect to a packaging process, a packaging material and / or a packaging type.

27. The method according to any one of claims 25-26, wherein the one or more reference values comprise one or more of: - a dimension of a goods receptacle, in particular a goods input opening, of the at least two packaging stations, and / or a shape, material type and / or property, strength, stability, surface quality of a packaging material of the at least two packaging stations, and / or - a dimension and / or shape of a final packaging and / or packaging raw material formed by the at least two packaging stations.

28. The method according to any one of claims 25-27, further comprising: - Determining, on the basis of the comparison, i) for which specific one of the plurality of packaging stations the individual product is suitable, and / or ii) how the individual product is to be handed over to the specific packaging station, in particular on the basis of the at least one property.

29. The method according to claim 28, wherein the comparing and / or determining steps i) and / or ii) comprises processing -- the image data, -- the generated individualized representation; -- the one or more reference values; and / or -- a feedback signal, in particular from one or more packaging stations and / or a cloud system, by means of the Kl system.

30. Method according to one of claims 19-29, wherein the control signal is configured to control a goods allocation device, in particular a robot, in particular a pick-and-place robot, in particular for picking up, in particular for gripping, the individual goods from the plurality of goods and / or for transferring the individual goods to the specific one of the plurality of packaging stations, in particular on the basis of the at least one property.

31. Goods allocation device, in particular a robot, in particular a pick-and-place robot, for packaging one or more goods, which implements the method according to one of claims 1-18 and / or the method according to one of claims 19-30.

32. A system for packaging one or more goods, comprising: - the goods allocation device according to claim 31, and - at least one image recording device, in particular with at least one camera, for obtaining the image data.

33. The system of claim 32, further comprising: at least one packaging station, or at least two packaging stations that differ in a packaging process, a packaging material, and / or a packaging type.

34. System according to any one of claims 33-34, further comprising: - at least one sensor device which is arranged -- to determine at least one further characteristic associated with the individual goods, and -- to provide at least one further property to the AI system as, in particular, additional, input.

35. System according to claim 34, wherein the at least one sensor device is located on the goods allocation device, in particular is attached to a goods receiving end point.

36. System according to one of claims 32-35, wherein the at least one sensor device comprises one or more of: at least one camera, at least one weight measuring device, in particular a scale and / or at least one laser measuring device.

37. The system of any of claims 32-36, further comprising: a cloud system communicatively coupled to the AI system, the cloud system comprising a database.

38. A computer program product comprising instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 1-18 and / or the method according to any one of claims 19-30.

39. A computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 1-18 and / or the method according to any one of claims 19-30.

40. Training data set for training an AI system and / or as input for the AI system, comprising: image data, in particular one or more images representing a plurality of goods with at least two, in particular different or identical, properties; and / or one or more generated individualized representation(s) of the plurality of goods; and / or at least one, in particular specific, property assigned to the individual good; and / or Data of the sensor device; one or more reference values for the, in particular predefined, packaging and / or for the corresponding goods receptacle, in particular the goods input opening, of the packaging station and / or for at least two of the plurality of packaging stations; and / or at least one feedback signal that is assigned to a product, in particular a specific property of the individual product; a product allocation device; a, in particular predefined, packaging, and / or a corresponding goods receptacle and / or at least one packaging station.

41. Generating a training data set according to claim 40 for training a Kl system.

42. Generating a training data set, in particular according to claim 41, for training a Kl system by means of the method according to one of claims 1-18 and / or the method according to one of claims 19-30 and / or the system according to one of claims 32-37.

43. Training an AI system using a training data set according to claim 40 and / or using the method according to any one of claims 1-18 and / or the method according to any one of claims 19-30.

44. AI system, in particular trained AI system which has been trained by means of the method according to one of claims 1-18 and / or the method according to one of claims 19-30 and / or the system according to one of claims 32-37; the training data set according to claim 40; and / or the training set generated according to any one of claims 41-42.

Images