• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The invention relates to an optimization method for improving the reliability of picking up and delivering goods in a picking method with a robot. A gripping unit on the robot head picks up goods from or from a first goods carrier and stores or discards them in or on a second goods carrier. In an image processing step, a gripping pose for the gripping unit is calculated for picking up the goods, in that at least one dimension is determined from a captured image and an area assignment is determined by comparison with dimension ranges. Depending on a confidence value, a dimension value is determined from the stored goods master data or from a standardization value of the dimension range and assigned to the determined dimension. In a preparation step, a mathematical scatter measurement function is used for the determined dimension and for the dimension ranges, and standardization values and confidence values of the dimension ranges are determined therefrom.
    公开号:AT521997A1
    申请号:T51021/2018
    申请日:2018-11-21
    公开日:2020-07-15
    发明作者:
    申请人:Tgw Logistics Group Gmbh；
    IPC主号:
    专利说明:

    a picking process with a robot and a work station.
    To increase picking performance, monotonous and tiring picking tasks are increasingly being carried out automatically by a robot system. However, in order that such a robot system can reliably process an order, in particular a picking order, it is necessary that relevant goods parameters are available with a high degree of reliability. The robot system here is understood to be the robot carrying out the goods manipulation with a gripping unit, and also the associated control. At least one dimension of the outer packaging is included here under goods parameters
    and / or understood the weight of the goods.
    When processing an order, a robot system requires information in advance about the number of goods to be transferred from the source goods carrier to the target goods carrier and how the goods are to be picked up from the source goods carrier (and, if applicable, how the goods are to be returned to the target goods carrier ). During picking, the robot system must check how many items still have to be transferred to fulfill the order. If there are goods at defined positions on or in the goods carrier, a robot can move to one of these positions and reliably pick up goods. However, if the goods are in a random, loose order, the first step is to determine where the goods to be picked up are located in / on the goods carrier. This is understood to mean explicitly that from a possible plurality of goods that can be taken up,
    the best possible goods to be picked is determined. Usually,
    N2018 / 30200-AT-00
    limit) that could not arise from a single item.
    For the correct processing of the order, it is also crucial that the correct number of goods is relocated. In the case of goods that are easy to grasp (goods with clear gripping positions), only one product is picked up by the gripper arm when the goods are picked up. However, goods are also to be picked which, for example, have an irregular external shape and in particular also polybags. Polybags are goods - often textiles - in an outer packaging made of a soft plastic film. With such a goods class, when the robot picks it up, it can happen that more than one item is picked up - which in any case must be recognized in order to prevent a picking error. The recognized goods are recognized, for example, by the weight of the goods.
    or target goods carrier is recorded.
    The reliability of picking therefore depends on goods parameters, which, due to their origin, can be subject to fluctuations. These parameters are usually determined manually when goods are received in the course of master data creation and are stored as goods master data in the warehouse management system (Warehouse Management System WMS), preferably
    together with other goods data such as an article number. There
    N2018 / 30200-AT-00
    Reduce the picking system.
    The present invention is based on the object of designing a picking method with a robot in such a way that the picking performance is independent of minor and significant fluctuations in goods parameters.
    becomes dependent.
    The object of the invention is achieved by an optimization method for improving the reliability of picking up and delivering goods in accordance with an order. The order is processed by a robot in a picking process, the robot having a robot head with a gripping unit that can be moved relative to a robot base. When the order is processed, a product is picked up by or from a first product carrier with the gripping unit and placed or discarded in or on a second product carrier. Such a goods carrier can be formed by a loading aid. Both goods carriers are preferably formed by loading aids. The loading aid is, for example, a container, box or tray. Therefore, the
    Goods were provided in a first loading device and in a second loading
    N2018 / 30200-AT-00
    of the gripping unit.
    According to the proposed solution, at least one dimension is determined in the image processing step when determining the gripping surface size and an area assignment is determined by comparing this dimension with dimension ranges. The dimension ranges are stored in a characteristic data field. A dimension value is determined for this area assignment and assigned to the determined dimension. This assignment takes place as a function of a confidence value of the dimension range, the dimension value being selected either from stored goods master data or from a standardization value of the dimension range. The normalization value and the confidence value
    of the dimension range are also stored in the goods characteristics field.
    Furthermore, in a preparation step, a mathematical scatter measurement function is used for the determined dimension and for the dimension ranges, and standardization values and confidence values of the dimension ranges are derived therefrom.
    averages.
    The particular advantage of this design is that the gripping surface size and the gripping surface pose can always be determined with great reliability. Goods to be picked may be subject to slight fluctuations in the dimensions of their outer packaging. Special packaging may also be used which, despite having the same article identification, has a different outer packaging - possibly even a significantly different outer packaging. Without the objective method, the dimensions can be incorrectly determined during image processing, so
    that the gripping surface size and / or the gripping surface pose is wrong or at least
    N2018 / 30200-AT-00
    working the order.
    The normalization value is understood to be the value that was determined in the preparation step from the dimensions determined in the image processing step. In a very simple case, this can be an arithmetic mean without restricting further explanations. In other embodiments according to the invention, other methods for forming the normalization value are provided. In order to determine the at least one dimension, an image of the goods in the first goods carrier is created by an optical image acquisition means which is connected to the opto-sensor processing and analysis system. For example, this image capturing means can be a CCD camera. According to a preferred embodiment, the image capturing means is formed by a 3D stereo camera which, in addition to the optical image, also supplies a point cloud of distance values. Several imaging devices can also be used
    the.
    In addition to the dimensions, the weight of the goods can also fluctuate. According to a further embodiment, it is therefore provided that the number of the picked up goods is determined in a counting step after the picking up of the goods by the gripping unit. In this case, a weight of the goods is determined by a weighing device, and this is compared with weight ranges and a range assignment is determined. The weight ranges are also stored in the characteristic data field. A number value is determined for this range assignment, this being determined as a function of a confidence value of the weight range, from stored goods master data or from a standardization value of the weight range. The normalization value and the confidence value of the weight range are also stored in the goods characteristics field. Subsequently, in the preparation step for the goods weight and for the weight ranges, a mathematical scatter measurement function is used, and therefrom standardization values and
    Confidence values of the weight ranges determined.
    N2018 / 30200-AT-00
    kidney drain is not or only slightly disturbed.
    The particular advantage of the method is that slight fluctuations in physical goods parameters such as external dimensions and weight, but also a (temporary) significant deviation from standard values, are recorded by the adaptive method. As a result, further processing can be adapted to the changed goods parameters and thus a faulty
    missioning can be avoided.
    A further development also consists in that, as an alternative to comparing the determined product weight with weight ranges, a quotient is formed from the product weight determined by the weighing device and a normalization value of the weight, and that this quotient is scaled and compared with weight ranges, and a range assignment is determined becomes. With this development, a simple possibility is advantageously created to draw direct conclusions from the determined goods weight on the number of goods picked up by the gripping unit. According to this version, the integer value corresponds directly to the number of goods received. By evaluating the decimal places, the spread of the determined weight and thus the reliability of the determined goods weight can be deduced. The smaller the difference between the value after the decimal point and the nearest integer value, the more precisely the determined weight of the goods corresponds to the normalized value of the weight or a multiple thereof. If the integer value is greater than one, the
    Quotient scaled to the weight range, which is the integer value of the quotient
    N2018 / 30200-AT-00
    decimal places.
    According to a further development, it is also provided that the weighing device is arranged separately from the robot, in particular the gripping unit, and detects the weight of the first goods carrier, and the counting step takes place before delivery of the goods, in particular after the goods have been picked up from the first goods carrier and before the gripping unit reached the second goods carrier. This embodiment has the advantage that the number of goods received is determined very shortly after the first goods carrier has been picked up. In this way, a reaction can be made quickly in the event of an incorrect pick-up, in particular before an incorrect delivery to the second goods carrier occurs. In a preferred embodiment, the weighing device is arranged such that the weight of the entire first goods carrier is determined. The goods weight recorded is therefore determined as a differential weight of the first goods carrier before and after the goods are removed. When the robot, especially the gripper arm, picks up the goods and moves away from the goods carrier, the goods accelerate. If the weighing device is arranged on the robot or on the gripper arm, this acceleration is recorded as an additional weight. A weighing device arranged separately from the robot and in particular from the gripping arm therefore has the advantage that a falsification of the goods weight due to the acceleration of the goods is prevented. It is therefore not necessary to correct the overlapping proportions of weight or to pause the robot to measure weight.
    A further development also consists in the fact that, in the case of a determined number value greater than one, a check is carried out as to whether the determined number value exceeds the target quantity of the goods to be picked, the target quantity being stored in the order and, if exceeded, a special treatment step is carried out. In the case of robot-assisted automatic picking, it is basically provided that only one piece of goods is picked up by the gripper arm and this process is carried out until the total number
    of the goods to be picked has been relocated. By arranging the
    N2018 / 30200-AT-00
    ß performed a special treatment step.
    It is also provided that a special treatment step is carried out if a limit value of the number value is exceeded or if a range limit value is exceeded when determining the range assignment. There can be situations where goods get caught or where several goods stick together and the gripping unit can pick up too many goods. According to a preferred embodiment, the method in question is designed to tolerate a mistake of up to three goods. It is explicitly pointed out that the method is not restricted to this number. If the determined goods weight is now larger than can be covered by the stored weight ranges, it is assumed that an undefined number of goods or an incorrect goods has been included. In this case, the picking process must not be continued, therefore it will be the
    Special treatment triggered.
    A further development consists in that the normalization value of the weight is formed from all normalization values of the weight ranges by means of a mathematical weighting function. By taking into account all standardization values of the weight ranges (in example 3) according to the formula
    "1 U2 HM3
    W = (Hı +> + 3) all individual standardization values (u) provide a weighted contribution to the determined weight of an individual item. It is explicitly pointed out that other mathematical weighting functions to form the standard
    weight can be used.
    N2018 / 30200-AT-00
    then lie in the first area and thus enable a clear assignment.
    A further development of the method in question is that the confidence values of the two adjacent weight ranges are analyzed for a third area, and the goods weight is determined from the normalization value of the weight area with the higher confidence value. The weight tolerances of the product weights can, for example, also overlap in such a way that the quotient or the fractional part lies exactly in the transition range of two weight ranges. According to this advantageous development
    the confidence values of the two weight ranges compared and the goods
    N2018 / 30200-AT-00
    weight determined on the basis of the weight range that has the higher confidence value. The confidence value provides information about the quality of the assignment of weight values to value ranges. The greater the confidence value, the more certain a specific determined product weight belongs to a weight range. This further training therefore increases the reliability of the determination of the goods weight, since an unlikely assignment due to the consideration of the
    Confidence value is excluded.
    To this end, it is also provided according to a further development that when determining the area assignment, a check is carried out to determine whether the determined dimension is in an overlap area of two dimension areas. If the test is positive, a special treatment step is carried out. The goods arranged in the first goods carrier can, for example, have an outer packaging in which at least two dimensional components differ only slightly from one another. In the image processing step it can therefore occur that a dimension is determined which lies in the transition area between two possible dimension ranges and is therefore not clearly assignable. In order to be able to conclude the order reliably, it is therefore provided in
    a special treatment step to resolve this ambiguity.
    After the special treatment has been formed, it is provided that, in the case of a determined position of the dimension in the overlap area, the image processing step is carried out again in the special treatment step. One way of resolving an ambiguous assignment of a determined dimension to a dimension range is to determine the dimension again. A new image is therefore preferably captured by an image capturing means and the individual steps of image recognition are carried out again. When the image is captured again, the lighting conditions may change, for example, and / or the image recognition will choose a different starting point, as a result of which the method steps of the image recognition come to a different, and possibly unambiguously assignable, result for the dimension. This step can also
    carried out again or several times, if necessary with activation
    N2018 / 30200-AT-00
    of illuminants to provide improved contrast for image recognition
    to accomplish.
    Another possible special treatment is that in the special treatment step, the robot is controlled to move the robot head to the first goods carrier, to pick up and put down a product, and that the image processing is then carried out again. In the event that no unambiguous dimension value can be determined in the image processing step, this development can provide for a more favorable starting situation for image recognition to be created by rearranging a product in the first product carrier. As a result of this rearrangement, the arrangement of the goods in the first goods carrier will change, as a result of which a newly recorded image is available for image processing. If necessary, this step can also be repeated
    be performed.
    Another possible development of the special treatment consists in that in the special treatment step the robot is actuated to move the robot head to the first product carrier, and to place or discard the picked up product on or in the first product carrier and that the image processing is carried out again. If it is not possible to clearly assign the determined weight to a weight range, it is advantageous with regard to a reliable execution of the order to put the received goods (or the goods) back into the first goods carrier and then to re-open the entire method start. The return of the goods received in the first goods carrier will result in a change in the arrangement of the goods in the goods carrier, which means a new one for the subsequent process steps (image processing, preparation and counting step)
    Starting point is created.
    According to a further development, it is further provided that in the preparation step, the normalization value for the assigned area is defined as the mean value of a normal distribution. This version is one way of using the determined values for the weight of the goods or for the determined dimension values to create a
    to get casual value for the respective parameter. The individual values of
    N2018 / 30200-AT-00
    Series of measurements of physical quantities such as weight and / or dimensions are usually subject to slight, random fluctuations. However, the individual values will accumulate around a medium value, larger deviations from this accumulation value will be rather rare. Such a measurement result can be interpreted as a normal distribution, for which a mathematical model exists for the calculation, with which the normalization value can be determined as the mean of the
    Normal distribution can be determined.
    An advantageous development consists in particular in that the scatter measure function comprises an expectation-maximization algorithm and iteratively combines determined values, dimensions or weight values to form local clusters. In a random process, as the determination of dimension values or weight values can be viewed, it can always happen that a determined value deviates significantly from the mean cluster value. In the present case, where there are several possible and valid dimensions or weight values, it can happen that a scattering value falls within the range of another accumulation value and would therefore be wrongly attributed to it. With the subject development, an iterative process is carried out which tries to combine the individual determined values into local clusters (clustering) in order to get a better result for the association of the measured values with local cluster values. Since the mathematical theory of this method is known per se, it is not described here
    discussed further.
    It is explicitly pointed out that for the description of the mathematical methods, the terms measurement value and determined value are synonymous.
    be applied.
    A further training course is that based on the summarized values
    a probability distribution is used, in particular a normal distribution, and that a mean value and a mean value scatter are determined for local clusters and the mean value is defined as a normalization value. On
    the local accumulations iteratively determined by the expectation-maximization algorithm
    N2018 / 30200-AT-00
    a normal distribution is used in each case in order to be determined by the iterative
    on to minimize the spread of mean values.
    In this respect, an advantageous further development is also characterized in that the scatter measure function comprises a sum of weighted normal distributions (Mixture of Gaussians) and iteratively combines determined values, dimensions or weight values to form local clusters. Weighted normal distributions, in particular mixed models, are a probability model for representing normally distributed partial results within a total set of results. In order to learn the model from the partial results, the expectation-maximization algorithm is used, which in an expectation step always assigns the partial results alternately to the various normal distributions and in a maximization step, based on the assignments made
    Optimized normal distribution parameters.
    A further development also consists in adapting the dimension ranges to the standardization values determined in the preparation step in an adaptation step, or in adapting the weight ranges to the standardization values determined in the preparation step. The subject method is designed as an adaptive method, so that each newly determined dimension and / or each newly determined product weight makes a contribution
    delivers to the respective dimensions or weight ranges.
    According to a further development, it is provided that for each dimension range and / or each weight range, the confidence value is determined from a spread of the normalization value, and / or that the confidence value is determined from the number of processing steps carried out. This further development makes it possible to state how well the individual determined value ranges reflect the actually occurring determined dimension and / or weight values. If the determined values are only very slightly scattered, a high confidence value will result, since the standardization values are then also subject to only very slight fluctuations. For example, in the case of a goods outer packaging with very unclear outer dimensions, it can happen that
    cannot clearly determine at least one goods limit in the image processing step,
    N2018 / 30200-AT-00
    and thus the values determined will be more scattered. In such a case, the confidence level will also be low. In order to stick to the example of the dimensions, it is also possible that only one dimension component is difficult to detect, but others are good. Since the confidence value for each weight range is determined after the development in question, the scattering dimension component and thus the assigned dimension range will have a small confidence value, another component with the assigned dimension range will have a higher confidence value. The number of processing steps carried out also has an influence on the confidence value. With an increasing number of processing steps carried out (with little variation), the certainty increases that the determined standardization value reflects the real situation. This means that the confidence value will increase with low scattering measurement values, but in return also that with strong ones
    scattering measured values the confidence value will decrease.
    According to a further embodiment, it is provided that a window function is applied to the determined dimensions and product weights when the standardization and confidence values are formed. The number of measured values considered in the preparation step is limited with a window function and thus a smooth determination of the standardization and confidence values is achieved. The sensitivity of the method can be set by the window width, in particular the reaction time to changing goods parameters is influenced. The smaller the window is selected, the faster the process in question reacts to changing measured values and can therefore adapt early, for example, to a changed outer packaging and / or to a changed amount of content (for example, if the manufacturer does the same
    A promotional product is brought onto the market).
    According to a further development, it is also provided that the goods master data are transmitted from a superordinate warehouse management system to the processing and analysis system and stored in a storage medium there
    become. Since the entire administration of the goods to be picked is
    orderly warehouse management system (WMS)
    N2018 / 30200-AT-00
    , it is advantageous if this also transmits the initial goods master data record to the processing and analysis system, as this saves work steps and avoids sources of error. In any case, this goods master data record includes a unique goods identifier, for example an article number of the WMS or an article code, for example an EAN. Characteristic or essential dimension data will also be included, for example length, width, height in the case of cuboid outer packaging or the smallest enveloping cuboid. Characteristic dimensions are understood to mean those dimensions which will preferably be optically detectable when the goods are arranged in the first goods carrier. The weight of the goods will also be stored in the goods master data record. The characteristic dimensions and weight are mostly
    nuelly recorded and entered into the WMS.
    An advantageous further development also consists in the fact that a sequence of goods transfers between the first and second goods carriers is carried out in the case of missing goods master data and the goods characteristic data field with standardization values and confidence values for dimensions and / or goods weights is built up from the respectively determined dimensions or goods weights. Most of the goods master data, including a clear goods identification, is provided by a warehouse management system (WMS). With the subject training, however, it is also possible to start building up your own goods characteristics without initial data from the WMS. For this purpose, the robot is controlled to pick up goods from the first goods carrier and to transfer them to the second goods carrier. The method steps according to the invention are carried out, in the preparation step the goods characteristic data field is set up and stored in the characteristic data field. In particular, the dimensional ranges are constructed with the respective standardization and confidence values and / or the weight ranges with the respective standardization and confidence values. These transfers are carried out a predefined number of times, so that a set of goods characteristics is subsequently available for the further course of the process. Based on a math
    Estimation or based on experience data, the number is chosen so that a
    N2018 / 30200-AT-00
    reliable characteristic data field can be built. With these rearrangements, it is in any case necessary that the goods in the first goods carrier are sorted
    are available.
    In order to accelerate the build-up of the characteristic data, according to a further development it is provided that a range of fluctuation is determined from a sequence of standardization or confidence values of the dimensions and / or of the goods weights, and the relocation of goods is terminated when the threshold of the fluctuation range is undershot. As suggested, the transfer process can be carried out as often as required in a certain number. However, it can also be checked to what extent the normalization value and / or the confidence value changes after each or after several rearrangement processes, and if there are only minor changes, the rearrangement process is terminated. At the beginning of the redistribution process, standardization and / or confidence values will change significantly, but then tend towards a value or level off around it. The longer the process is carried out, the smaller the fluctuations will be, so the range of fluctuation will be smaller. According to the development in question, the rearrangement can now be terminated if the threshold of the fluctuation range is undershot, since the accuracy achieved is sufficient for the actual picking process. During picking, the characteristic data are continuously adapted in the processing step, in particular
    improves.
    According to a further development, it is also provided that a complete relocation is carried out and the goods carrier from which it is relocated is completely emptied. After the rearrangement, the normalization values and the confidence values determined are analyzed. If a sufficiently high data quality could be achieved through this rearrangement process, which is reflected in a correspondingly high confidence value, the initial characteristic data formation step can be ended. If necessary, a complete migration can also consist of a complete transfer of a source product to a target product and then the target product back to the
    Source goods carrier is relocated. It is also possible that during the
    N2018 / 30200-AT-00
    Relocation of the normalization values and the confidence values are analyzed, and if the data quality is sufficient, the relocation process is reversed so that the goods that have already been relocated are returned to the outgoing
    goods carriers are moved.
    According to an advantageous further development, it is also provided that in the case of a high confidence value for a dimension standardization value or for a weight standardization value, an automated or partially automated update of the goods master data is carried out. As already described, goods master data, as provided by the warehouse management system (WMS), can deviate from the real dimensions or weight values, which is often not recorded in the goods receipt process. Therefore, the values (dimensions and / or weight) determined with the method in question will no longer match the values provided by the WMS. However, the respective method will adapt the respective standardization values to the newly determined values, and consequently the confidence value will also increase. From a threshold value of the confidence value, the normalization values determined with the method in question have a higher reliability than the data from the WMS. In particular, this also means that the goods master data must be incorrect, as otherwise there should not be a significant deviation. In the case of an automated update of the goods master data, the goods characteristic data determined using the method in question are transferred directly to the WMS and the goods master data is then updated from this. In the case of a partially automated update, a message is sent to the WMS or to a responsible supervisor, with the actual update being triggered manually
    must become.
    The object of the invention is also achieved by a work station, in particular a picking station. The work station comprises a delivery device for one or more goods carriers, a robot with a robot head that can be moved relative to a robot base, and a gripping unit for transferring
    tion of goods between product carriers, an image capture device, and a
    N2018 / 30200-AT-00
    Data processing unit designed to control the robot (and possibly the conveyor technology) and also to evaluate data from the image capture device. It is provided that the data processing unit is designed to carry out an objective optimization process. The advantage of this embodiment is that a device for carrying out the method in question can be very easily integrated in a picking system. It is particularly advantageous that the workstation in question is also suitable for exchanging or expanding existing picking systems. This means that existing order picking systems can also be moved
    automatic picking can be expanded.
    In one possible embodiment, the robot is used at a picking station and the goods are picked up from or from a first goods carrier and placed or discarded in or onto a second goods carrier by means of the gripping unit, and the optimization method according to the invention is carried out in the process. However, the robot can also be used at a packing station, where the goods for an order (without reference to a specific picking order) are picked up from or from a first goods carrier in the described manner and placed or discarded in or on a second goods carrier, and the optimization method according to the invention executed. In other words, the robot can be used at any workstation equipped with a robot in order to carry out the optimization method according to the invention. The work station can comprise a data processing unit designed to control the robot (and possibly the conveyor technology) and also to evaluate data from the image capture device. On the other hand, the data processing unit can be formed on a central computer unit. The data processing unit is used to implement the inventive
    optimization process.
    The delivery device can be formed by a parking space on a storage table or a conveyor system for the delivery and removal of goods carriers. The goods handling device is formed by a robot, for example
    se an articulated arm robot or a portal robot.
    N2018 / 30200-AT-00
    Below is a definition of terms used herein.
    A robot is understood here to mean automatic machines with multiple axes, in particular articulated arm and portal robots. Both systems are designed to position a robot head within a coordinate space relative to a robot base and to perform manipulation actions at the specified coordinate with a gripping unit arranged on the robot head. The positioning of the robot head, in the coordinate space defined by the mechanical structure of the robot, is usually carried out by means of a program
    of a numerical control system.
    An order is recorded electronically as a data record and processed by the data processing unit. The order is not necessarily a picking order, but can also be a repacking order, replenishment order, and the like. A picking order is understood to be the compilation of requested goods from a customer order. A customer order includes at least one order. Each order comprises one or more order lines. If the order specifies several order lines, different goods are required. Lot sizes in e-commerce are small, so that there is a relatively high number of different orders, each with a few order lines. We speak of a repacking order if, for example, goods are to be reloaded from a delivery container (first goods carrier) to one or to a second goods carrier in the incoming goods department. A replenishment order is the filling up of goods
    in or into a product carrier.
    Commodity or SKU (stock keeping unit) is understood to mean the smallest commissionable goods packaging. This unit can consist of a single article, but it is also possible that a product comprises several articles, which are, however, not separated for picking. It is pointed out that the normalization value and the confidence value for the weight and dimensions are fundamentally comparable, and essentially
    allow the same statement. Accordingly, the descriptions are
    N2018 / 30200-AT-00
    to understand individual advantages in principle on both value classes, except when
    explicitly directed to dimensions or weight.
    In the warehouse management system (WMS), characteristic goods parameters are preferably recorded in the goods receipt process and stored as goods master data in a storage medium of the WMS. This goods master data can be transmitted to the optimization process in question in the course of initial initialization and can be stored, for example, in a storage medium of the processing and analysis system. The parameters determined by the possible method and possibly deviating are stored as product characteristics, for example also in the memory.
    stored in the processing and analysis system.
    The preparation and analysis system is to be understood as the control system of the system for carrying out the method in question. In particular, this also includes the robot controller, which, based on the determined coordinates of the gripping pose, moves the robot head with the gripping unit to the given position in the coordinate space. The robot controller can be integrated in the preparation and analysis system, but it is also possible to run it independently with a connection via an interface. For the
    The objective method is the integrated consideration.
    A mathematical scatter measurement function is understood here to mean a mathematical calculation model with which (mostly minor) variations in the measurement data acquisition can be classified. In particular, this is intended to ensure that a recorded measured value can be assigned to a unique data value. Through the iterative application of these
    on the reliability of the assignment can be increased.
    It is also explicitly pointed out that the descriptions of mathematical models or determination methods are applicable to the treatment of determined dimensions and determined goods weights, unless it is explicitly for an application to a determined dimension value or to a determined
    The weight of the goods is limited. Equally important is also
    N2018 / 30200-AT-00
    indicated that the process steps in question are intended for the treatment of dimension and weight variations, unless explicitly
    zit restricted to one option.
    Depending on the design of the product carrier, the product can be arranged in the product carrier if it is designed as a container, for example. In the case of training as a shelf, the goods will be arranged on the goods carrier. It is explicitly pointed out that both terms are used synonymously here and none by the use of in or on the goods carrier
    Definition of a specific training of the goods carrier is made.
    The dimensions of a product are understood to be the (outer) dimensions of the outer packaging. The outer packaging can be a box, for example, then the dimensions are the lengths of the respective edges. An outer packaging can also be formed by a shrink film, which adapts to the contour of the goods and thus has no straight outer contour. In this case, the dimensions can be formed, for example, by straight line interpolations of the outer contour. These examples are not conclusive
    possible variants such as product dimensions are to be understood.
    For a better understanding of the invention, this is based on the following
    Figures explained in more detail.
    1 shows an embodiment of the system and of the method in question to make variations in the dimensions of the outer packaging
    being able to grasp and treat;
    2 shows an embodiment of the system and of the method in question in order to make variations in the dimensions of the outer packaging
    being able to grasp and treat;
    3 shows a representation of value ranges as they are formed with the method in question and for the determination of distance values or
    Number values are used.
    N2018 / 30200-AT-00
    Fig. 1 shows an embodiment of the subject method to improve the reliability of picking up and delivering goods to detect fluctuations in the dimensions of the outer packaging and in determining the
    Grasping pose to be able to take into account.
    At an automatic picking station 1, goods are made available in or on goods carriers 3 by means of a conveyor technology 2 - after the picking, the goods carrier or goods carriers 3 are removed again by the conveyor technology 2
    and preferably the next product carrier (s) 3 is provided.
    The picking station 1 forms a possible embodiment of a work station, which has a delivery device for one or more goods carriers 3, a robot 4 with a robot head that can be moved relative to a robot base, with a gripping unit 5 for transferring goods between goods carriers 3, an image capture device, and one for controlling the Robot (and possibly the conveyor technology) and further include data processing unit designed for evaluating data from the image capture device. According to this version, the delivery device is on conveyor system 2 for
    and / or removal of goods carriers 3 is formed.
    The provision and the removal are not essential for the process in question, so it will not be discussed in more detail. The goods carrier 3 can be formed, for example, by containers or trays without restricting the specific embodiment. The conveyor technology 2 can also be carried out without restriction of the specific design, for example by roller or belt conveying.
    be educated.
    A robot 4 has a gripping unit 5 on a robot head 6, which robot head 6 can be moved relative to a robot base 7 and can be moved by a control unit 8 into any gripping position within the range of motion of the robot 4. The range of motion of the robot in any case includes the area in which the goods carriers 3 (source and target) are arranged. Thus, the gripping unit 5 can reach any area of the goods carrier 3 in order to
    pick up or drop off goods there.
    N2018 / 30200-AT-00
    To determine the gripping pose, at least the source charge carrier 3 is captured by an image acquisition system 9 of an opto-sensor processing and analysis system 10 of the control unit 8 in a first step. If the goods to be picked have to be delivered to the target load carrier 3 at a specific position, the target product carrier is also captured by another image acquisition system 9. The image acquisition system 9 is preferably designed as a stereo camera system and, in addition to an optical image, also records a point cloud with distance data between the camera system 9 and the surface of the goods on the goods carrier 3 (and the edges of the goods carrier 3). The captured image 11 and the determined point cloud are transmitted to the control unit 8 and there are processed by the opto-sensor processing and analysis system.
    tem 10 prepared and evaluated.
    1, the captured image 11 is shown in a simplified manner, in particular without the captured edge of the product carrier 3 and with a vertical view of the product carrier 3. However, this is of no significance for the method steps to be described. A chaotically arranged goods 12 can be seen in the image 11 of the goods carrier 3. In addition to individually arranged goods, it will always happen, particularly for goods carriers with a higher degree of filling, that goods are also arranged one above the other. To simplify the perspective
    Display of stacked goods incomplete.
    In order to be able to pick up the goods 12 and to hold them securely during the transfer into the target goods carrier, the gripping unit 5 must, to put it simply, pick up the goods as centrally as possible in the gripping surface 13. The gripping unit 5 can be equipped, for example, with gripping arms or preferably with at least one suction element, it also being of great importance for reliable pick-up that the robot 4 positions the gripping unit 5 as normally as possible on the gripping surface 13. One advantage of receiving the goods 12 as centrally as possible is that the goods are then usually also picked up near the center of gravity, which is advantageous for a secure hold during the relocation movement. If the goods 12, for example, by flexible
    Polybags is formed, a central recording as possible has the further advantage
    N2018 / 30200-AT-00
    that the goods then hang approximately equally far down on both sides of the gripper in the direction of the goods carrier. When relocating, the goods must be lifted by the robot 4 to such an extent that they do not touch any existing edges during the swiveling movement from the source goods carrier to the target goods carrier
    of the goods carrier 3, still at elements of the picking station 1.
    By means of an edge detection method of the opto-sensory processing and analysis system 10, product boundaries are searched in the captured image 11 of the product carrier 3, which is possible, for example, by means of a contrast detection method. It can now happen, for example, that goods 12 are randomly arranged 14 in such a way that a supposedly continuous edge is created 15. The entire length 15 would then be recognized as a dimension by the contrast detection method, for example, and based on this, an incorrect gripping pose would be calculated. With the consequence that it is very likely that the goods cannot be picked up or will detach from the gripping unit during the relocation. In Fig. 1 a simplified situation is shown where only two goods are close to each other or one above the other. Furthermore, the possibility of delimitation based on the distance data determined by the image acquisition system 9 was not taken into account for this description. In a real, chaotically filled goods carrier, a large number of goods will be arranged close to one another and one above the other, and will also not be reliably distinguishable using the distance data (especially in the case of very flat polybags), so that the probability of recognizing an inadmissible edge length and thus determining a invalid gripping pose will rise. Ultimately, such picking errors will make the picking performance
    decrease or the frequency of errors increases.
    The security of the detection of an individual piece of goods is increased according to the method in question by comparing the dimensions determined with goods master data that were recorded in the goods receipt process. As already described, these dimensions can fluctuate. To grasp a valid gripping pose - reliably to grip a single item
    to be able - must now have a determined dimension with stored dimensions
    N2018 / 30200-AT-00
    areas are matched, that is, an assignment of the stored dimension values to the determined dimensions. As long as the individual dimensions of the outer dimensions are clearly different, a comparison with the goods master data will certainly give good results. Temporary changes to the outer packaging by the manufacturer, with the same goods number, cannot be made using a procedure based purely on goods master data
    are recorded and will lead to picking errors.
    With the method in question, the dimensions determined in the image 11 are compared with an adaptively adapted value distribution 16 in order to be able to compensate for fluctuations in the dimensions. Measured values that are subject to random fluctuations will usually have a normal distribution of the values. A basically cuboid-shaped piece of goods has three dimension coordinates, the majority of the detected dimension values being each distributed around one of the three dimension normalization values 17. Details on the properties of normal distributions are not elaborated on here, since these are known to a person skilled in the art. Since a normal distribution describes a possible distribution function, the general terms normalization value and distribution curve are used below, the normalization value in
    Special case of a normal distribution corresponds to the mean.
    The edge detection method of the opto-sensory processing and analysis system 10, for example, recognized the outline for a product and used it to determine two dimension values X1, X2. In order to determine which of the three possible dimensions of the outer packaging is the determined dimension value X1 18, the determined value X1 is projected onto the value distribution 16. It can be seen that the determined value X1 is close to the normalization value u3 17 and within the distribution curve 19 belonging to this normalization value 17. According to the method according to the invention, there is a confidence value C for each distribution curve 19, which represents a measure of how well the normalization value 17 is supported by the distribution curve 19. A more detailed description follows below in the text. Since the determined dimension
    value X1 is close to the normalization value u3 and the confidence value C3 for this
    N2018 / 30200-AT-00
    Distribution curve N3 is high, the dimension value becomes equal to the standard
    value set Uu3.
    As soon as the confidence value C is above a definable threshold value, the normalization value determined is selected as the dimension value in accordance with the method in question; below this threshold, the dimension value is selected from the goods master data. Since the method in question adapts the normalization value and the confidence value each time dimension values are determined, assuming slightly scattering measured values, with increasing frequency of execution, the certainty that the correct dimension values have been determined will increase. Since the goods master data is not updated, there will be a discrepancy between the goods master data and the standardization values if, as already described, an outer packaging is changed. In this case, however, due to a possibly high confidence value, the values determined using the method in question are used and the gripping surfaces are thus despite different goods packaging.
    size and the gripping pose correctly determined.
    In addition to the three dimension ranges u1-M3, a fourth range with a distribution curve 28 is also shown. Values that fall into this range cannot be assigned to a real dimension of a product, since they can only originate from an entry error or an inadmissible product that is erroneously located in / on the product carrier. In this case, the image processing step should preferably be carried out again in a special treatment. If such incorrect recordings occur frequently, information could be derived, for example, that goods of this type should be arranged differently in the goods carrier, if necessary,
    to prevent detection errors.
    To reliably process an order, in addition to reliably picking up and delivering goods, it is also necessary that the correct number of goods be transferred. According to a further embodiment of the method illustrated in FIG. 2, the weight of the goods picked up by the gripping unit 5 is determined to determine the number of goods picked up.
    rated.
    N2018 / 30200-AT-00
    The picking station 1 differs from the picking station described in FIG. 1 essentially in that a weighing device 20 is arranged in the area of the source goods carrier. The weighing device 20 can also be arranged in the region of the target goods carrier or on the robot head 6. The arrangement in the area of the source goods carrier has the advantage that the number of goods picked up can be determined immediately after the goods have been picked up, in particular before the robot arm 21 moves in the direction of the target goods carrier. This makes it possible to correct a mistake very early, for example if more goods are taken up than are required to fulfill the order. However, the inclusion of several goods is permitted as long as the total number of goods to be picked
    sioning goods is not exceeded.
    Due to a difficult surface for the image analysis (for example glossy and / or strong contrasts), it can happen that the gripping pose has not been optimally determined and the gripping unit has been positioned just close to the edge of the goods, for example. When goods are picked up, it can happen that several goods are picked up. The change in weight of the source goods carrier is recorded by the weighing device 20 and evaluated using the method according to the invention in order to determine the reliable number of
    to determine goods.
    Just like the outer dimensions, the weight of the goods can also be subject to slight fluctuations or, as already described for the outer packaging, can be temporarily changed by the manufacturer. The basic problem is therefore equivalent to the situation when determining the outer dimensions for determining the gripping surface size or gripping surface pose, so that this description
    exercise is not repeated here.
    As for the dimensions, the one captured by the weighing device 20
    Weight change W1 22 projected onto the value distribution 23. It can be seen that the determined value W1 is close to the normalization value u1 24, and the weight of the piece of goods picked up is slightly less than the standard
    mation value u1 24. In this case, with a confidence value above the
    N2018 / 30200-AT-00
    Provided that the threshold value can be reliably determined that exactly one piece of goods has been picked up, the number value can therefore be set at one
    become.
    For example, a different weight can be detected by the weighing device 20, which after projection onto the value distribution 23 lies in the range between the normalization values u2 and u3 and lies above the normalization value u2. Again, depending on the confidence value, the number value is set to 2 in this case. If, according to the order, at least two goods are still to be relocated, the control unit 8 will move the robot head 6 to the target goods carrier and place the goods there. If, however, only one piece of goods needs to be picked, the controller 8 will control the gripping unit 5 to return the picked-up goods to the source goods carrier. If necessary, it can be provided that the goods are not handed in at the place of the pick-up, but somewhat offset from them in order to change the grasping pose for the next determination
    and possibly create a more favorable starting position.
    FIG. 3 shows a representation of how a determined value (a dimension or a weight) is projected onto the value distribution and the dimension value or the number value is determined above it. Since this classification and assignment procedure for dimensions and product weight is basically the same, the following
    understand the description explicitly for both parameters.
    FIG. 3 shows a value distribution 26 with three distribution curves 19, which each describe the distribution of measured values recorded by a normalization value 17. This distribution of values 26 arises adaptively after several runs of the method in question, since in the preparation step a scatter measurement function is used for each of the measured values recorded, and standardization values and confidence values are determined therefrom. During the first run of the method in question, the value distribution will not exist, only with repeated runs, and thus repeated use of the scatter measurement function on the measured values recorded, (stable) normalization values 17 and distribution curves 19 will form. In particular, with increasing
    Number of adaptation runs including the confidence values of the individual standards
    N2018 / 30200-AT-00
    values rise. However, it does not necessarily mean that all confidence values have to increase. For example, a measured value can vary widely despite a large number of preparation steps, so there will be a wide distribution curve
    set and the confidence level will therefore be low.
    After several runs of the process in question, a
    Set the value distribution 26 shown in a playful manner.
    After the determination of a dimension or after the determination of a product weight (generally a measurement value), the objective method compares the dimension or the product weight with areas of the value distribution. In particular, the measured value is projected onto the value ranges 27 and a range assignment is determined. In the value distribution 26 in FIG. 3, several value ranges are shown, which are due to the adaptive execution of the subject
    process.
    There is a region (I, Ill, V) around each normalization value 17, where an unambiguous assignment of the measured value to a normalization value is possible. If the confidence value is also above a definable threshold, the normalization value can be assigned to the dimension value or the number value.
    the.
    Between the individual distribution curves 19 there are transition or overlap areas (II, IV) in which a clear assignment of a distance value or a number value is not possible. If the measured value lies in one of these ranges, the method according to the invention provides that special treatment is carried out. As a special treatment, it can be provided, for example, that the two adjacent distribution functions are analyzed and, in particular, the two confidence values are used to decide which of the two standardization values is selected. In simplified form, the scatter and the number of measured values are included in the determination of the confidence value, so that the normalization value with the higher confidence value is usually selected. This also results in the asymmetrical position of the transition areas
    che. Without taking the confidence value into account, the allocation limit
    N2018 / 30200-AT-00
    are exactly at the intersection of the two distribution curves 19, or a possible transition area would be significantly wider. The advantageous method according to the invention can now be used to shift both the position, that is to say the width, of the transition region in favor of the distribution curve with the higher confidence value. If a previous assignment of a standardization value to a measured value subsequently led to a picking error, the history of the previous assignments can also be
    be taken into account in order to increase the quality of the decision.
    The figure also shows two edge areas (A, B), for which error handling must also be carried out if a measured value is assigned to this area. These areas essentially represent incorrectly recorded measured values. The special treatment for these cases usually consists in the fact that the image acquisition is carried out again or the goods are returned from the gripping unit to the goods carrier. A measured value in one of these areas can also mean that an unintended item is on the item carrier. In the special treatment, a higher-level warehouse management system can be alerted and the goods carrier can be
    nik be transported to a manual workplace.
    The arrangement of the goods carriers 3 shown in relation to the robot 4 is only to be understood as an example, and was chosen in this way, in particular for reasons of illustration. In any case, there are also other configurations
    possible.
    In conclusion, it should be noted that in the differently described embodiments, the same parts are provided with the same reference numerals or the same component designations, the disclosures contained in the entire description being able to be applied analogously to the same parts with the same reference numerals or the same component designations. The location information selected in the description, e.g. above, below, laterally etc. related to the figure described and illustrated immediately and are at one
    Transfer the change of position to the new position.
    N2018 / 30200-AT-00
    The exemplary embodiments show possible design variants of the, it being noted at this point that the invention is not restricted to the specially illustrated design variants of the same or the same, but rather also various combinations of the individual design variants with one another are possible and this variation possibility is based on the teaching of technical action due to the subject invention lies in the ability of the person skilled in the art in this technical field. So all conceivable design variants, which are possible by combining individual details of the illustrated and described design variant, are also protected by
    catch with includes.
    For the sake of order, it should finally be pointed out that for a better understanding of the procedural steps, drawing elements are sometimes not to scale
    and / or enlarged and / or reduced.
    N2018 / 30200-AT-00
    32
    Reference list
    Picking station conveyor technology
    Goods carrier
    robot
    Gripping unit
    Robot head
    Robot base
    Control unit image acquisition system opto-sensory processing and analysis system image
    Would
    Gripping surface
    arrangement
    dimension
    Distribution of values for dimensions
    dimension
    Weighing device distribution curve
    Robotic arm
    Weight
    Distribution of goods weight standardization value
    Weight
    Distribution of values
    Value ranges
    Distribution curve
    N2018 / 30200-A T-00
    权利要求:
    Claims (25)
    [1]
    1. Optimization method for improving the reliability of picking up and delivering goods according to an order in a picking process with a robot, which robot has a robot head with a gripping unit that can be moved relative to a robot base, in which goods are picked up with or by a first goods carrier with the gripping unit and deposited or discarded in or on a second goods carrier
    will, and
    a gripping surface size and a gripping surface pose are determined in an image processing step using an opto-sensor processing and analysis system, and a gripping pose for the gripping unit is calculated therefrom, the gripping unit is moved into the calculated gripping pose and the goods are gripped by the gripping unit
    is recorded, characterized in that in the image processing step
    e When determining the gripping surface size, at least one dimension is determined
    will and
    e A range assignment is determined by comparing this dimension with dimension ranges, the dimension ranges being measured in a
    ren characteristic data field are stored, and
    e a dimension value is determined for this area assignment and is assigned to the determined dimension, the dimension value being selected from a stored master data of the goods or from a standardization value of the dimension area depending on a confidence value of the dimension area, the standardization value and the confidence value of the dimension area
    are stored in the goods data field, and that
    in one preparation step
    N2018 / 30200-AT-00
    e a mathematical scatter measurement function is used for the determined dimension and for the dimension ranges, and therefrom standardization values
    and confidence values of the dimension ranges are determined.
    [2]
    2. Optimization method according to claim 1, characterized in that in one, the receiving of the goods downstream by the gripping unit
    Counting step, the number of goods picked up is determined by determining a weight of the goods, and
    e the goods weight with weight ranges, which are stored in the characteristic data field
    are compared and an area assignment is determined, and
    e for this range assignment, a number value is determined, depending on a confidence value of the weight range, determined from stored goods master data or determined from a standardization value of the weight range, the standardization value and the confidence value of the weight range being stored in the goods characteristic data field,
    and that in the preparation step
    e for the goods weight and for the weight ranges, a mathematical scatter measurement function is used, and therefrom standardization values and
    Confidence values of the weight ranges can be determined.
    [3]
    3. Optimization method according to claim 2, characterized in that, as an alternative to the comparison of the determined goods weight with weight ranges,
    chen,
    e from the product weight determined by the weighing device and a standard
    weighting a quotient is formed, and that
    e the quotient is scaled and compared with weight ranges, and a loading
    empire assignment is determined.
    N2018 / 30200-AT-00
    [4]
    4. Optimization method according to claim 2 or 3, characterized in that the weighing device is arranged separately from the robot, in particular the gripping unit, and detects the weight of the first goods carrier, and the counting step takes place before delivery of the goods, in particular after the goods have been picked up from the first Goods carrier and before the gripping unit the second goods carrier
    reached.
    [5]
    5. Optimization method according to one of claims 2 to 4, characterized
    characterized in that in the case of a determined number value greater than one, a check is made as to whether the determined number value exceeds the target quantity of the goods to be picked, the target quantity being stored in the order, and at
    Exceeding, a special treatment step is carried out.
    [6]
    6. Optimization method according to one of claims 2 to 4, characterized in that in the event that a limit value of the number value is exceeded, or if a range limit value is exceeded when determining the
    Area assignment, a special treatment step is carried out.
    [7]
    7. Optimization method according to one of claims 2 to 6, characterized in that the normalization value of the weight by means of a mathematical weighting function from all normalization values of the weight ranges
    is formed.
    [8]
    8. Optimization method according to one of claims 3 to 7, characterized
    characterized in that a range classification is carried out for the decimal value of the quotient and the number value is determined by the integer part of the quotient for a first range, and one for a second range
    Special treatment is carried out.
    [9]
    9. Optimization method according to claim 8, characterized in that
    that for a third area the confidence values of the two adjacent areas
    N2018 / 30200-AT-00
    weight ranges are analyzed, and the goods weight from the standardization
    value of the weight range with the higher confidence value is determined.
    [10]
    10. Optimization method according to one of claims 1 to 9, characterized in that a test is carried out for the determined dimension when determining the area assignment,
    whether the ascertained dimension is in an overlap area of two dimension areas and, in the case of a positive test, a special treatment
    step is carried out.
    [11]
    11. Optimization method according to claim 10, characterized in that in the case of a determined position of the dimension in the overlap area, in the special
    the treatment step the image processing step is carried out again.
    [12]
    12. Optimization method according to claim 10 or 11, characterized in that the robot is controlled in the special treatment step
    To move the robot head to the first product carrier, and to pick up and put down a product and that the image processing is then carried out again.
    to be led.
    [13]
    13. Optimization method according to one of claims 5 to 12, characterized in that in the special treatment step, the robot is controlled to move the robot head to the first product carrier, and the recorded one
    Place or discard goods on or in the first goods carrier.
    [14]
    14. Optimization method according to one of claims 1 to 13, characterized in that in the preparation step, for the assigned area, the
    The normalization value is determined as the average of a normal distribution.
    [15]
    15. Optimization method according to one of claims 1 to 14, characterized
    characterized that the scatter measure function is an expectation-maximation
    N2018 / 30200-AT-00
    Algorithm includes and determined values, dimensions or weight values,
    summarized iteratively to local clusters.
    [16]
    16. The optimization method as claimed in claim 15, characterized in that a probability distribution, in particular a normal distribution, is applied to the combined values, and that a mean value and a mean value scatter are determined for local accumulations and the mean value as
    Normalization value is set.
    [17]
    17. Optimization method according to one of claims 1 to 15, characterized in that the scatter measure function comprises a sum of weighted normal distributions (Mixture of Gaussians) and determined values, dimensions
    gen or weight values, iteratively summarized to local clusters.
    [18]
    18. Optimization method according to one of claims 1 or 3 to 17, or one of claims 2 to 17, characterized in that in an adaptation step the dimension ranges are adapted to the normalization values determined in the preparation step, or that
    in the adaptation step, the weight ranges to those determined in the preparation step
    partial standardization values are adjusted.
    [19]
    19. Optimization method according to one of claims 1 to 18, characterized in that for each dimension range and / or each weight range, the confidence value is determined from a spread of the normalization value, and / or that the confidence value is based on the number of treatments performed.
    steps are determined.
    [20]
    20. Optimization method according to one of claims 1 to 19, characterized in that when the standardization and confidence value are formed, a window function depends in each case on the dimensions and goods weights determined.
    is turned.
    N2018 / 30200-AT-00
    [21]
    21. Optimization method according to one of claims 1 to 20, characterized in that the goods master data are transmitted from a superordinate warehouse management system to the preparation and analysis system
    and stored in a storage medium.
    [22]
    22. Optimization method according to one of claims 1 to 21, characterized in that in the case of missing goods master data, a sequence of goods relocations is carried out between the first and second goods carriers, and from the respectively determined dimensions or goods weights the goods characteristic data field with standardization values and confidence values for dimensions
    and / or goods weights are built up.
    [23]
    23. Optimization method according to claim 22, characterized in that a fluctuation range is determined from a sequence of standardization or confidence values of the dimensions and / or the product weights, and the rearrangement of the goods is less than a threshold of the fluctuation range.
    will end.
    [24]
    24. Optimization method according to one of claims 1 to 23, characterized in that, with a high confidence value for a dimension standardization value or for a weight standardization value, an automated one
    or semi-automated updating of the goods master data is carried out.
    [25]
    25. work station, in particular picking station, comprehensive
    a delivery device for one or more goods carriers,
    a robot with a robot head that can be moved relative to a robot base with a gripping unit for relocating goods between goods carriers,
    an image capturing device,
    and a data processing unit designed to control the robot (and possibly the conveyor technology) and also to evaluate data from the image capture device,
    characterized in that
    N2018 / 30200-AT-00
    the data processing unit for performing a method according to one of the
    Claims 1 to 24 is formed.
    N2018 / 30200-AT-00
    类似技术:
    公开号 | 公开日 | 专利标题
    DE102015115115B4|2019-03-21|An article receiving device for picking up randomly heaped objects
    AT507339B1|2011-06-15|METHOD AND APPARATUS FOR RECORDING AND / OR PROCESSING OBJECTS
    EP2984007B1|2017-03-29|Storage and order-picking system for the fully-automated identification and picking of articles, and corresponding order-picking station
    DE112011103794B4|2019-01-24|Pick-up device for workpieces
    EP2686254B1|2018-08-15|Controlling and monitoring a storage and order-picking system by means of movement and speech
    DE102014102943A1|2015-08-13|Robot system with functionality for determining the location of a 3D box
    AT519452B1|2018-07-15|Method and device for picking goods
    DE112020000009T5|2020-10-01|Robotic system with coordination mechanism and method for operating this system
    DE202011004401U1|2012-06-18|Control and monitoring of a storage and picking system by movement and language
    AT521997A1|2020-07-15|Optimization process to improve the reliability of goods picking with a robot
    DE102020100610A1|2020-07-16|CONTROL AND CONTROL METHOD FOR ROBOTIC SYSTEM
    EP2818433B1|2016-08-10|Automated picking cell and method for the automated picking of A+ articles
    DE102016120809B4|2020-12-24|Method for the robot-assisted fitting of a male component into a female component
    DE102015225587A1|2017-06-22|Interaction system and method for interaction between a person and at least one robot unit
    EP3854731A1|2021-07-28|Device and method for separated piece goods
    WO2019197611A1|2019-10-17|Picking up payloads by means of a robot-guided tool
    DE102018208126A1|2019-11-28|Method for handling a workpiece with the aid of a removal tool and machine for carrying out the method
    WO2016165694A1|2016-10-20|Method and device for storing objects in a shelf system
    WO2020083586A1|2020-04-30|Method for creating an object model for gripping an object, computer-readable storage medium, and robot system
    DE112017007399B4|2021-09-23|Control device, gripping system, distribution system, program, control method and manufacturing method
    EP2452291A1|2012-05-16|Method for continuously determining a gripping position
    EP3838806A1|2021-06-23|Fully automated robotic picking system and picking method with sensor assembly for measuring a removed item as part of the lifting process
    DE102017002354A1|2018-09-13|Transporting objects, in particular picking goods, using a robot
    DE102010051935A1|2012-05-24|Method for removing component from component carrier, involves moving gripping tool towards component stack until gripping tool of robot is contacted with highest component of component stack, so as to seize components
    DE102017216863B3|2018-12-27|Method and robot system for adapting a robotic insert to a work process
    同族专利:
    公开号 | 公开日
    AT521997B1|2021-11-15|
    US20210402595A1|2021-12-30|
    CA3118595A1|2020-05-28|
    EP3884349A1|2021-09-29|
    WO2020102840A1|2020-05-28|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5041907A|1990-01-29|1991-08-20|Technistar Corporation|Automated assembly and packaging system|
    US20180215545A1|2017-01-30|2018-08-02|Wal-Mart Stores, Inc.|Systems and Methods for Resolving Issues in a Distributed Autonomous Robot System|
    US9561587B2|2014-12-16|2017-02-07|Amazon Technologies, Inc.|Robotic grasping of items in inventory system|
    CN109983433A|2016-07-18|2019-07-05|L·奥德纳|Evaluate robot crawl|DE102020210537A1|2020-08-19|2022-02-24|Kuka Deutschland Gmbh|Method and system for handling a load assembly with a robotic gripper|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    ATA51021/2018A|AT521997B1|2018-11-21|2018-11-21|Optimization process to improve the reliability of goods picking with a robot|ATA51021/2018A| AT521997B1|2018-11-21|2018-11-21|Optimization process to improve the reliability of goods picking with a robot|
    PCT/AT2019/060395| WO2020102840A1|2018-11-21|2019-11-20|Optimization method for improving the reliability of goods commissioning using a robot|
    CA3118595A| CA3118595A1|2018-11-21|2019-11-20|Optimization method for improving the reliability of article picking using a robot|
    US17/289,887| US20210402595A1|2018-11-21|2019-11-20|Optimization method for improving the reliability of goods commissioning using a robot|
    EP19828185.9A| EP3884349A1|2018-11-21|2019-11-20|Optimization method for improving the reliability of goods commissioning using a robot|
    [返回顶部]