Method and device for dividing up a packaged product by means of a robotic arm
Patent Information
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- ROBOMINDS GMBH
- Filing Date
- 2024-06-04
- Publication Date
- 2026-04-22
AI Technical Summary
Existing methods for separating packaged goods using robot arms often result in damage due to inappropriate grip force, especially when the goods are unknown or unidentifiable, leading to inefficiencies and potential damage during handling.
A method involving sensor detection of packaged goods to create a measured value matrix, followed by inference using a statistical model to determine the extent, coverage state, position, and orientation, which generates a control data set for the robot arm to perform a controlled pulling movement for separation, reducing grip force and preventing damage.
This approach allows for reliable, safe, and efficient separation of packaged goods without damage, enabling quick and cost-effective handling while minimizing space requirements.
Smart Images
Figure EP2024065275_19122024_PF_FP_ABST
Abstract
Description
[0001] Method and device for separating a packaged product by means of a robot arm
[0002] The invention relates to a method for separating packaged goods by means of at least one robot arm according to claim 1 and to a device comprising a robot arm for separating packaged goods according to claim 10.
[0003] Robots are known in the art which, using a robot arm and a suitable gripper, pick up packaged goods from a pile of packaged goods, lift them up and move them to a predetermined storage location, and place them there (so-called pick & place robots), where they can be processed individually and separately from the other goods. It is important that the robot's access and the subsequent movements carried out with the object are adapted to the packaged goods in such a way that the goods are not damaged. In particular, too great or too great a gripping force can damage the packaged goods, as they are crushed or fall down. The way in which the packaged goods must be held so that no damage occurs is highly object-specific.In particular, if the packaged goods are postal items, identifying the packaged goods is undesirable and / or illegal. However, this poses the problem that, for example, the weight and / or nature of the packaged goods are also unknown. Therefore, there is always a risk of damage.
[0004] There is therefore a great need for a method and device for separating packaged goods, where the packaged goods are preferably not known and / or identified beforehand. The handling of the packaged goods should be as reliable, safe, consistent, and fast as possible without delay, slowing down, or interruption. In addition, the method should be cost-effective to carry out, and the device should be cost-effective to manufacture and operate, with a further focus on making them as compact and space-saving as possible. The object of the invention is to provide such a method and device, overcoming the disadvantages of the prior art while requiring the smallest possible space.
[0005] This object is achieved in a surprisingly simple but effective manner by a method for separating packaged goods according to the teaching of the independent main claim 1 and a device comprising a robot arm for separating packaged goods according to the teaching of claim 10.
[0006] According to the invention, a method for separating packaged goods by means of at least one robot arm is proposed, the method comprising the following steps: a. detecting at least one measured value matrix of packaged goods by means of at least one sensor; b. determining an extent, a coverage state, a position and / or an orientation of at least one specific item of the packaged goods by means of inference with a statistical model and / or by means of an approximation method with input of the measured value matrix; c. creating a control data set for the robot arm taking into account the extent, the coverage state, the position and / or the orientation of the specific item; d. grasping the specific item by means of the robot arm according to the control data set; and e.Separating the specific goods from the other packaged goods using at least one pulling and separating movement according to the control data set by means of the robot arm.
[0007] Within the scope of the invention, it has been recognized that the separation of unknown and packaged goods by lifting regularly fails because the packaging and / or the packaged goods are damaged either by excessive gripping force of the gripper arranged on the robot arm or by falling due to insufficient gripping force. Within the scope of the invention, it has further been recognized that the gripping force can be significantly reduced if the specific good is not lifted, but is separated by means of a pulling movement directed away from the other packaged goods on a surface on which the packaged goods are placed. Within the scope of the invention, it has been recognized that the movement is preferably an optimized movement in the pulling direction. On the one hand, this requires less gripping force than for lifting the specific good.On the other hand, if the gripping force is too low, the specific item will not fall, but will simply remain in its current position. The gripper can then re-grasp the specific item and continue the process. Furthermore, it has been recognized that separation can be achieved even with short pulling distances, and in particular, gentle setting down of the specific object is not necessary, making the process quick and effective, while also saving space.
[0008] In an initial situation, at least two packaged goods are arranged in a heap-like arrangement. Preferably, at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100 identical and / or different packaged goods are arranged in a heap-like arrangement. The terms “packaged goods” and “specific goods” are understandable to a person skilled in the art and refer, for example but by no means exclusively, to a parcel, a small package, a postal item, a consignment of documents, a letter, luggage such as a suitcase and a bag, a crate, a carton, a box and / or bulky goods.
[0009] In order for the packaged goods to be separated at all, the packaged goods must first be detected in step a. The detection of the packaged, transported goods takes place by measuring and creating a measured value matrix using at least one sensor. It is obvious to a person skilled in the art that the packaged goods are detected partially or completely. It is conceivable that the measured values for the measured value matrix are detected by the sensor simultaneously or sequentially. Furthermore, it is conceivable that the sensor is part of a measuring device with a data processing device and that the measuring device creates the measured value matrix by calculating the individual measured values determined by the sensor, in particular while the sensor is moving. The measured values of at least 2,
[0010] 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 similar and / or different sensors simultaneously or successively. More preferably, at least 2, 3,
[0011] 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 identical or different measured values are measured. Even more preferably, at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 sensors record at least one measured value.
[0012] The term “separation” or “separation” refers to the removal and / or separation of the specific goods from the other packaged goods so that they are isolated from the other packaged goods.
[0013] The term “measured value” refers to the quantization, expression and / or digitization of a property of the arrangement of the packaged goods.
[0014] The term "measurement matrix" refers to a data structure used to store and organize measurement values. The data structure consists of a two-dimensional array in which the measurement values are arranged in rows and columns. The two dimensions correspond to the spatial directions of the surface on which the packaged goods are arranged, whereby each spatial point in the two dimensions can be assigned one or more measurement values.
[0015] In step b., the measured value matrix created in step a. is entered into a statistical model and / or determined using an approximation method. The statistical model and / or the approximation method was trained before and / or is trained during the execution of the method using training data sets. The training data sets can be created simulatively and / or consist of real process data. The statistical model infers and / or the approximation method determines an extent, a coverage state, a position and / or an orientation of at least one specific item. Preferably, the statistical model and / or the approximation method interferes with the extent, the coverage state, the position and / or the orientation of all packaged goods recorded in the measured value matrix. The specific item can be selected randomly from the recorded items in order to enable the fastest possible handling.The statistical model is preferably a statistical data model that is generated using an approximation method, in particular deep learning, convolutional neural network (CNN), recursive nets (RNN), generative adversarial networks (GAN), stochastic machine, random forest or support vector machine.
[0016] The term “dimension” refers to the geometric shape of the packaged goods.
[0017] The term “coverage status” refers to the degree of coverage of a packaged good which is covered or covered by one or more other packaged goods.
[0018] The term "position" refers to the location and / or orientation of a packaged good relative to a reference point and / or a coordinate system. The reference point and / or coordinate system can refer to the surface on which the packaged goods are arranged, the position of the robot arm, or other, particularly neighboring, packaged goods.
[0019] The term ‘orientation’ refers to the orientation of the packaged goods in relation to the surface on which the packaged goods are arranged, taking into account the dimensions of the packaged goods.
[0020] In step c., a control data set for the robot arm is created, taking into account the coverage state, the position, and / or the orientation of the specific item. Particularly preferably, the control data set is created such that the direction of subsequent movement is determined by the position in the direction of the shortest distance to a surface edge, in the direction of the lowest coverage state, and / or in the direction of the lowest degree of tilt determined from the orientation. Particularly preferably, the control data set is created by inference with the statistical model or by inference with another statistical model and / or by the approximation method. The statistical model and / or the approximation method was trained before and / or is trained during the execution of the method using training data sets with regard to the control data set.Furthermore, a predetermined final position of the specific item can be taken into account when creating the control data set. The training data sets preferably comprise exemplary dimensions, coverage states, positions, and / or orientations of packaged goods in heap-like arrangements of packaged goods and associated optimal and / or correct control data sets. Even more preferably, an optimal and / or correct control data set is created by inference. It is conceivable that the statistical model and / or the approximation method infers the at least one, preferably optimal and / or correct, control data set from the fact that the extent, coverage state, position, and / or orientation of the packaged item has the highest correspondence with the exemplary packaged item in a training file.Even more preferably, it is conceivable for the statistical model and / or the approximation method to infer the at least one, preferably optimal and / or correct, handling property from the fact that the extent, the coverage state, the position and / or the orientation of the packaged goods have the highest agreement with individual training data sets and the control data sets stored for the individual training measured values are combined to form an optimal and / or correct control data set. In other words, the inference of the statistical model and / or the approximation method corresponds to the at least one control data set of a specific training data set, or the control data set is formed by inference from various control data sets of different training data sets. The advantage of the first approach is that the control data set is guaranteed to be logically structured.The advantage of the second approach is that the control data set is more individual and therefore potentially more optimal and / or more accurate for the current situation. In addition to the extent, coverage state, position and / or orientation of the specific item, it is also conceivable that further information outside the specific item is taken into account. The information can preferably relate to the robot arm, in particular a status of the robot arm, the environment in which the method is carried out, in particular a status of the environment, and / or the user, in particular specifications of the user using the method. The control data set specifies the movement sequences that the robot arm is to carry out.Within the scope of the invention, at least one robot arm is provided, preferably 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 similar and / or different robot arms arranged next to one another or one behind the other.
[0021] The term ‘optimal and / or correct control data set’ refers to a control data set which a reasonable and knowledgeable person skilled in the art, who recognises the heap-like arrangement of the packaged goods and in particular the extent, the state of coverage, the position and / or the orientation of the specific goods, would classify as optimal and / or correct in the given environment in which the process is carried out.
[0022] In step d, the robot arm grasps the specific item according to the control data set. This means that the robot arm picks up the specific item using the gripper.
[0023] In step e., the robot arm separates the specific item from the other packaged goods using at least one pulling and separating movement. This means that the specific item is pulled away from the other specific goods in a lateral movement. Since the specific item is not lifted, it is possible to reduce the gripping force and effectively prevent it from falling. Furthermore, the stress on the packaging of the packaged goods is also reduced, so that it is less exposed to the risk of tearing. It is conceivable to perform the process several times in succession or simultaneously with several robot arms, each with different specific goods, until some or all of the packaged goods have been separated.
[0024] The method according to the invention can be repeated any number of times and can be carried out fully or partially automated. It has been recognized that each specific product undergoes the steps essential to the invention, with these steps depending on the specific product.
[0025] The method according to the invention enables a robotic arm to automatically separate a specific item in a simple, fast, and easily adaptable manner. The method is cost-effective to implement and safe in handling and packaging the goods. The method is therefore reliable, fast, and safe, as well as space-saving.
[0026] Advantageous further developments of the invention, which can be implemented individually or in combination, are presented in the subclaims.
[0027] It is conceivable that in step e., the movement takes place essentially horizontally. In this way, it is possible to drastically reduce the tensile and / or pulling force acting on the packaged goods, in particular by reducing the weight of the goods by partially or completely resting them on the surface on which the packaged goods are placed. This also reduces the stress on the packaging of the packaged goods, so that they are less exposed to the risk of tearing.
[0028] The term "substantial" means that there is only a minor, in particular insignificant, change, alteration, and / or deviation from the relevant conditions. With regard to movement, "substantially horizontal" means that the minor, in particular insignificant, deviation from horizontal movement, such as due to the different height and / or orientation of the surfaces on which the packaged goods are placed, a conveyor belt, and / or a continuous conveyor, is insignificant and is therefore compensated.
[0029] In an advantageous development of the method, packaged goods and / or the specific good are transported before, during and / or after steps a. to e. Within the scope of the invention, it has been recognized that the separation serves in particular to distribute the packaged goods. The separated packaged goods are to be brought to different locations so that further transport of the separated goods is achieved after the method. Removal of the separated goods is also advantageous so that further separated goods can then be brought to the released location and the separated movement can be kept to a minimum. Transport before the method enables a constant supply of goods to be separated, so that the method can be carried out without interruption and with a correspondingly high throughput.Therefore, further time savings can be achieved if the procedure is carried out before, during or after the transport of the specific goods and / or the packaged goods.
[0030] In a further development, it is conceivable that the number of packaged goods transported in and the number of packaged goods removed from the system, particularly after separation, are measured. If the number of packaged goods transported in is greater than the number of packaged goods removed, the transport of the packaged goods, particularly the inward transport, is interrupted, the transport of the packaged goods, particularly the outward transport, is increased, and / or the inward transport and outward transport are coordinated. If the number of packaged goods removed is lower than the number of packaged goods transported in, congestion occurs, and the process can no longer be carried out safely and reliably.If the inbound transport is temporarily or briefly interrupted, the outbound transport is temporarily or briefly increased, and / or the inbound and outbound transport are coordinated, the congestion can be resolved using the procedure. Transport can then be resumed. A specialist will be familiar with suitable procedures and measures for temporarily or briefly interrupting transport, increasing transport, and / or coordinating these.
[0031] Furthermore, it is advantageously conceivable for the robot arm to be moved at a speed relative to the transport speed of the packaged goods. This achieves a coordination of the transport speed and the movement speed of the packaged goods, thereby minimizing the load on the packaged goods or the packaging of the packaged goods. It is conceivable for the relative speed to affect the entire robot arm. This means that the robot arm is transported together with the packaged goods, at least in sections. Alternatively, only the gripper can be moved at the relative speed with the packaged goods, which is made possible by appropriately coordinated movements of the rest of the robot arm.
[0032] It is also advantageously conceivable for the robot arm to rotate the specific item by a rotation angle and / or tilt it by a degree of tilt before and / or during step e. in a substantially horizontal direction. This reduces the contact area between the specific item and the surface on which the packaged goods are arranged, and a reduced frictional force is created between the specific item and the surface when the pulling, substantially horizontal and separating movement in step e. is carried out. The load on the specific item or the packaging of the specific item is therefore further reduced and can also be determined and / or adjusted. The angle of rotation is preferably at least 1°, 2°, 3°, 4°, 5°, 6°, 7°, 8°, 9°, 10°, 12.5°, 15°, 17.5°, 20°, 25°, 30°, 35°, 40° or 45°.More preferably, the degree of tilt is at least 1°, 2°, 3°, 4°, 5°, 6°, 7°, 8°, 9°, 10°, 11°, 12°, 13°, 14°, 15°, 16°, 17°, 18°, 19°, 20°, 25°, 30°, 35°, 40° or 45°. Even more preferably, the specific item is tilted by the degree of tilting and lifted by at least 0.5 mm, preferably at least 1 mm, 2 mm, 2 mm, 3 mm, 4 mm, 5 mm, 6 mm, 7 mm, 8 mm, 9 mm or 10 mm, or a maximum of 10 mm, preferably a maximum of 9 mm, 8 mm, 7 mm, 6 mm, 5 mm, 4 mm, 3 mm, 2 mm, 1 mm or 0.5 mm. In this way, the frictional force generated in step e., in particular the frictional force on the edge of the item located on the surface, is further reduced and can be determined and / or adjusted in such a way that the load on the specific item or the packaging of the specific item is therefore further reduced.This is because the weight of the goods is reduced, which is particularly useful for large, heavy, coated, sticky, wet and / or film-wrapped goods due to their increased adhesion to the surface on which the packaged goods are placed.
[0033] In a further development, it is conceivable that the specific item is rotated one or more times before and / or during step e. This enables the item to be identified by positioning it through one or more rotations around the longitudinal and / or transverse axis in such a way that the identification, inscription, marking, and / or labeling is legible and / or recognizable.
[0034] Furthermore, it is advantageously conceivable that by querying the measured value matrix, a section of the measured value matrix and / or another measured value from another sensor in a database and / or inference with the and / or another statistical model and / or the approximation method, a handling property of the specific item is determined, wherein the handling property is a target address, a grip strength, a point of attack, an orientation, a selection of a gripper, a pulling speed and / or a pulling strength. Within the scope of the invention, it has been recognized that, depending on the type of packaged item and / or the packaging of the packaged item, different handling properties are conceivable. In particular, for goods packaged in bags, a greater grip strength and / or a different gripper is required than for goods packaged in boxes.The grippers available for selection preferably comprise a suction gripper, a magnet and / or a mechanical gripper, in particular a pincer gripper, a three-finger gripper and / or a four-finger gripper. Furthermore, one, two, several or all grippers can be arranged on the robot arm simultaneously, and / or a change of grippers may be possible. In this way, the gripping movement and / or the gripper can be optimally adapted to the specific product and the load can be kept to a minimum. It is conceivable that the measured value matrix, the section of the measured value matrix and / or the further measured value comprises a readable, in particular a machine-readable, indication, in particular a code. A barcode and / or a QR code is preferably conceivable.
[0035] Preferably, it is conceivable for the steps of the method to be carried out continuously or in a time-discrete manner. A continuous execution enables high throughput and therefore significant time savings, while a time-discrete execution offers high process reliability. It is conceivable for some of the steps to be carried out continuously and others in a time-discrete manner. Furthermore, it is conceivable for a choice to be made between a continuous and a discrete mode. In this way, the method can be adapted to the packaged goods present, the environment, and / or an initial situation.
[0036] Furthermore, it is advantageously conceivable that in step b. the extent, the coverage state, the position and / or the orientation of at least two detected goods from the packaged goods are determined by means of inference with the and / or another statistical model and / or the approximation method upon input of the measured value matrix, and a specific good is selected from the packaged goods. Within the scope of the invention, it has been recognized that, given the heap-like arrangement of the packaged goods in the situation presented, one or more packaged goods are more suitable for carrying out the method according to the invention on them than others due to their extent, their coverage state, their position and / or orientation. This can arise, for example, from the fact that the more suitable packaged goods are located closer to an edge of the heap-like arrangement than the others.Therefore, it is efficient to first carry out the method according to the invention on these objects. Preferably, the extent, coverage, position, and / or orientation of all objects recorded in the measured value matrix are determined by inference with the and / or another statistical model and / or the approximation method, inputting the measured value matrix.
[0037] It is also advantageously conceivable for the method to include a further step f., namely evaluating the separation and obtaining an evaluation, as well as optionally displaying the evaluation. This evaluation makes it possible to assess the quality of the separation. Advantageously, it is also possible to provide training data sets created from the evaluation for expanding and / or updating the statistical model and / or the approximation method. In this way, the separation can be continuously optimized during operation.In particular, the statistical model and / or the approximation method can be continuously trained, improved, and / or adapted through the evaluation, so that for new situations not contained in the training data, the optimal and / or correct handling characteristic can be inferred with a higher probability using the static model upon repeated repetition of the method according to the invention. It is conceivable to perform the evaluation using the existing sensors and / or using sensors installed specifically for evaluation.
[0038] It is assumed that the definitions and descriptions of the above terms apply to all aspects described below in this description, unless otherwise stated.
[0039] According to the invention, a device for handling objects comprising a robot arm, at least one sensor, and a computer system is further proposed, wherein the computer system comprises a computer-readable storage medium comprising a statistical model and / or an approximation method for executing step b of a method described elsewhere and preferably executed by the device, and at least one data processing device for executing steps b, c, d, and / or e of the method described elsewhere. By means of the device according to the invention, the advantageous method can be carried out, along with the advantages described there. Particularly preferably, the device comprises at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 identical or different sensors and / or 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 identical or different robot arms.Even more preferably, the computer-readable storage medium comprises the further statistical model(s) described elsewhere and / or the database described elsewhere and / or the approximation method described elsewhere. The device according to the invention is fully or partially automated.
[0040] It is also conceivable for the sensor to be part of an image capture device, in particular a 3D camera, a lidar, a radar, a tactile sensor, a spectrometer, an ultrasound measuring device, an X-ray device, a distance measuring device and / or a refractometer. In principle, it is arbitrary what type of sensor is used to measure the measured value, as long as it satisfies the claims according to the invention and can record a measured value suitable according to the invention for a measured value matrix suitable according to the invention. Which object property is suitable according to the invention depends to a large extent on the intended use of the device. The aforementioned sensors and / or devices are particularly suitable for one or more specific uses. The corresponding uses and the sensors and / or devices suitable for them will be apparent to those skilled in the art.
[0041] Furthermore, it is conceivable for the device to comprise a further sensor, wherein the further sensor is a light sensor, a force-torque sensor, a vacuum sensor, a force sensor, a rotary encoder, a momentum sensor or a pressure sensor or is part of a device, in particular a barcode scanner, an RFID scanner, or a measuring device, in particular a thermometer, an explosives measuring device, a hazard measuring device, a moisture measuring device and / or a scale. The further sensor is suitable for detecting the further measured value described elsewhere for determining at least one handling property in order to achieve the advantages described there. The further sensor can also be arranged on and / or in the gripper and evaluate the movement. In particular, the further sensor can detect whether the gripper is losing the specific item. In this case, a correction of the control data set is possible.The sensor can also detect explosives and / or hazardous substances and separate packaged goods that contain explosives and / or hazardous substances.
[0042] In a further development of the invention, it is conceivable for the robot arm to be single-axis or multi-axis. A single-axis robot arm has the advantage that it is easier to control and monitor. A multi-axis robot arm, on the other hand, can cover a larger area of reachable packaged goods and, in particular, grasp and separate them in a more correct and / or optimal manner. The robot arm is particularly preferably guided on a rail. Furthermore, it is conceivable for the robot arm to comprise at least one, preferably 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 identical or different suction grippers, a magnet and / or a mechanical gripper, in particular a pincer gripper, a three-finger gripper and / or a four-finger gripper.
[0043] Furthermore, it is conceivable for the device to comprise at least one continuous conveyor and / or a load carrier. In this way, the transport of the packaged goods before, during and / or after steps a. to e., as described elsewhere, can be achieved, along with the advantages described therein. More preferably, the device comprises at least 2, 3, 4, 5, 6, 7, 8, 9 or 10 continuous conveyors running at the same and / or different speeds and / or in the same and / or different directions. It is preferably conceivable for the specific goods to be pulled onto a predetermined continuous conveyor during singulation. The load carrier particularly preferably has an openable and / or open side wall.
[0044] In a further development of the invention, the device comprises a staging surface. The packaged goods are arranged on the staging surface in a heaped arrangement for carrying out the method. The staging surface is particularly preferably designed to have the lowest possible static friction and high abrasion and / or scratch resistance. The staging surface is particularly preferably made of stainless steel.
[0045] It is further conceivable for the computer system to comprise an interface and / or a network interface. The network can be an intranet comprising at least one other computer system and / or the internet. The interface and / or network interface make it possible to make training data sets and / or database entries available to the computer system for creation. Entries and / or updates to the database, in particular handling properties, can also be made using interface and / or the network interface. It is further conceivable for the database to be located partially or completely outside the computer system and for the computer system to be connected to the database via the network interface. The manner in which the entries and / or updates for the database are generated is fundamentally arbitrary.It is preferably conceivable that the entries and / or updates are made using the measured values recorded by the computer system, as described elsewhere. Even more preferably, it is conceivable that the database can be updated and / or expanded by at least one other identical or different computer system, which is also capable of recording measured data, and / or by storage during production and / or recording of the packaged goods. In particular, this can be the database described elsewhere for general networking and digital accessibility. Furthermore, it is possible to provide training data sets for training the statistical model and / or the approximation method using the interface components and / or the Internet interface. This can enable safe, correct and / or optimal handling of individual objects in advance.Particularly preferably, destination addresses can be passed on to the computer system by means of the interface parts and / or a network interface.
[0046] In a further development of the invention, it is conceivable that the device is a parcel sorting system and the packaged goods, as described in detail elsewhere, are, for example, mail items. In principle, the type of packaged goods to which the method according to the invention is applied is arbitrary. However, mail items are packaged goods that regularly have to be separated for further transport. They are also differently shaped, have different packaging, their contents are unknown, and they have different destination addresses. Therefore, the advantages achievable with the method and / or the device according to the invention are particularly suitable for a parcel sorting system.
[0047] Further proposed according to the invention is a computer program product comprising software code sections configured such that a method described elsewhere can be executed by at least one processor. This achieves the advantages described elsewhere.
[0048] Further details, features, and advantages of the invention will become apparent from the following description of the preferred embodiment in conjunction with the subclaims. The respective features can be implemented individually or in combination with one another. The invention is not limited to the embodiment. The embodiment is shown schematically in the figure.
[0049] In detail:
[0050] Fig. 1 is a schematic side view of a device according to the invention.
[0051] Fig. 1 shows a schematic side view of a device 21 according to the invention. The device 21 is a parcel sorting system. The device comprises a robot arm 13, at the end of which a gripper is arranged. In Fig. 1, the robot arm is designed with two axes. On the left of the image, several differently designed packaged goods 10 are arranged in a pile. The packaged goods 10 are mail items. The packaged goods 10 lie on a first continuous conveyor 11, which transports the packaged goods 10 into the area of a sensor 17. The sensor 17 is designed as a 3D camera. The sensor 17 detects the packaged goods 10 and creates a measured value matrix, wherein the measured value matrix is a 3D image. From the measured value matrix, an extent, a coverage state, a position and / or an orientation of all detectable packaged goods 10 is first created using a statistical model executed on a computer system 15.These are the packaged goods 10 that can be detected by the sensor 17 because they are at least partially uncovered. From this, a specific item 14 is interfered with by the statistical model. In the present case, this is the packaged item 10 located on the right in Fig. 1 because it is closest to the gripper. From the extent, the coverage state, the position and / or the orientation of the specific item 14, the statistical model infers a control data set, which is transmitted to a control system 16 for controlling the robot arm 13. The gripper grasps the packaged item 14 and initially rotates it through a rotation angle 18 relative to the essentially horizontal surface of the continuous conveyor 11. The robot arm 13 then moves the gripper and thus the specific item 14 in a pulling, essentially horizontal and separating movement 19 according to the control data set, without lifting the specific item 14.The movement 19 ends on a second continuous conveyor 12, which transports the specific goods 14 together with other packaged goods 20 to be transported away.
Claims
Patent claims 1. Method for separating packaged goods (10, 14, 20) by means of at least one robot arm (13), comprising the following steps: a. Creating a measured value matrix of the packaged goods (10, 14) by means of at least one sensor (17); b. Determining an extent, a coverage state, a position and / or an orientation of at least one specific item (14) of the packaged goods (10, 14, 20) by means of inference with a statistical model and / or by means of an approximation method with input of the measured value matrix; c. Creating a control data set for the robot arm (13) taking into account the extent, the coverage state, the position and / or the orientation of the specific item (14); d. Gripping the specific item (14) by means of the robot arm (13) according to the control data set; and e.Separating the specific goods (14) from the other packaged goods (10) by at least one pulling and separating movement (19) according to the control data set by means of the robot arm (13).
2. Method according to claim 1, wherein in step e. the movement (19) is substantially horizontal.
3. Method according to claim 1 or 2, wherein the packaged goods (10, 20) and / or the specific good (14) are transported before, during and / or after steps a. to e.
4. Method according to claim 3, wherein the number of transported packaged goods (10) and the number of transported packaged goods (20) are measured and, in the event that the number of transported packaged goods (10) is greater than the number of transported packaged goods (20), the transport of the packaged goods (10), in particular the transport in, is interrupted, the transport of the packaged goods (10), in particular the transport out, is increased and / or the transport in and the transport out are coordinated with one another.
5. Method according to claim 3 or 4, wherein the robot arm (13) is moved at a speed relative to the transport speed of the packaged goods (10, 20).
6. Method according to one of the preceding claims, wherein the specific good (14) is rotated on the substantially horizontal by a rotation angle (18) and / or tilted by a tilting degree before and / or during step e.
7. Method according to one of the preceding claims, wherein the specific good (14) is rotated one or more times before and / or during step e.
8. Method according to one of the preceding claims, wherein by means of querying the measured value matrix, a section of the measured value matrix and / or a further measured value of a further sensor in a database and / or inference with the and / or a further statistical model at least one handling property of the specific good (14) is determined, wherein the handling property comprises a target address, a grip strength, an attack point, an orientation, a selection of a gripper, a pulling speed and / or a pulling strength.
9. Method according to one of the preceding claims, wherein the steps of the method are carried out continuously and / or time-discretely.
10. Method according to one of the preceding claims, wherein in step b . the extent, the covering state, the position and / or the orientation of at least two detected goods ( 14 ) from the packaged goods ( 10 , 14 ) is determined by means of inference with the and / or a further statistical model by inputting the measured value matrix and a specific good ( 14 ) is selected from the packaged goods ( 10 , 14 ). 1 1. Method according to one of the preceding claims, wherein the method comprises a further step f.: evaluating the singulation and obtaining an evaluation.
12. Device (21) for separating packaged goods (10, 14, 20), comprising at least one robot arm (13), at least one sensor (17) and a computer system (15), characterized in that the device (21) is configured to carry out the method according to one of claims 1 to 11, wherein the computer system (15) comprises a computer-readable storage medium comprising the statistical model for carrying out step b. and at least one data processing device for carrying out step b., c., d. and / or e.
13. Device (21) according to claim 12, characterized in that the sensor (17) is part of an image capture device, in particular a 3D camera, a lidar, a radar, a tactile sensor, a spectrometer, an ultrasonic measuring device, an X-ray device, a distance measuring device and / or a refractometer.
14. Device (21) according to claim 12 or 13, characterized in that the device (21) comprises a further sensor, wherein the further sensor is a light sensor, a force-moment sensor, a vacuum sensor, a force sensor, a rotary encoder, a momentum sensor or a pressure sensor or is part of a device, in particular a bar code scanner, an RFID scanner, or a measuring device, in particular a thermometer, an explosives measuring device, a hazard measuring device, a moisture measuring device and / or a scale.
15. Device (21) according to one of claims 12 to 14, characterized in that the robot arm (13) is single-axis or multi-axis.
16. Device (21) according to one of claims 12 to 15, characterized in that the device (21) comprises at least one continuous conveyor (11, 12) and / or a load carrier.
17. Device (21) according to one of claims 12 to 16, characterized in that the device (21) comprises a provision surface.
18. Device (21) according to one of claims 12 to 17, characterized in that the computer system (40) comprises an interface module (41) and / or a network interface (42).
19. Device (21) according to one of claims 12 to 18, characterized in that the device (21) is a parcel sorting system, and the packaged goods are parcels, packages, mail, document mail, letter mail, luggage, crates, cartons, boxes, and / or bulky goods.
20. A computer program product comprising software code sections designed such that a method according to one of method claims 1 to 11 can be executed by at least one processor.