Systems and methods for sku induction, pour, and automated eligibility estimation

Abstract

An object induction system for assigning handling parameters to objects is disclosed. The system includes an analysis system, a correlation system, and an assignment system. The analysis system includes at least one property sensing system for providing sensing data regarding objects to be processed. The correlation system includes an object information database and assigns correlation data to the objects in response to commonalities of any of the property sensing data to any of property record data. The assignment system is for assigning programmable motion device handling parameters to the marker sensing data based on the correlation data and includes a workflow management system and a separate operable controller.

Description

[0001] priority

[0002] This application claims priority to U.S. Provisional Patent Application No. 63 / 107,680, filed October 30, 2020, the entire disclosure of which is incorporated herein by reference. Background Technology

[0003] The present invention generally relates to programmable motion processing systems, and more specifically to programmable motion (e.g., robots) systems designed for use in environments, for example, where multiple objects (e.g., articles, parcels, or packages) need to be processed (e.g., sorted and / or otherwise distributed) to several output destinations.

[0004] Many object distribution systems receive objects in organized or unorganized flows, which can be provided as individual objects or in groups (such as in bags) to reach any of several different modes of transport, typically conveyors, trucks, pallets, gaylords, or bins. Each object must then be distributed to the correct destination location, as determined by identification information associated with that object, which can be identified by a label printed on the object. The destination location can take various forms, such as bags, bins, or crates.

[0005] An automated system for order fulfillment was also proposed. However, the challenge is that any automated system must be able to handle a wide variety of objects of different sizes, weights, volumes, centers of mass, and shapes in many applications, as well as objects and / or packages with low placement permissions (such as clothes packaged in plastic bags) or low positioning permissions (such as round or cylindrical objects that can be rolled or otherwise moved after being placed).

[0006] Furthermore, many object distribution systems receive objects (e.g., SKU items, parcels, packages, etc.) from one or more shipping entities for distribution to a wide variety of destinations. Therefore, such object distribution systems sometimes must consider the generalization of objects for which little or no information is available. While such unknown objects can be handled manually, in automated object handling systems, such handling by humans may be destructive or otherwise impractical.

[0007] The object distribution system still needs a more efficient and cost-effective object sorting system that distributes objects of various sizes and weights into appropriately sized collection bins or pallets, but is efficient in handling such objects of different sizes, weights, compositions, and labels. Summary of the Invention

[0008] According to one aspect, the present invention provides an object classification system for assigning handling parameters to objects. The object classification system includes an analysis system, an association system, and an assignment system. The analysis system includes at least one feature sensing system for providing sensing data about the object to be processed. The feature sensing data includes data related to any one of weight, height, width, length, weight, center of mass, object description, object category, and at least one image. The association system includes an object information database having feature recording data, which includes data related to any one of weight, height, width, length, weight, center of mass, object description, object category, and at least one image of multiple objects. The association system is used to assign association data to objects in response to commonalities between any of the feature sensing data and any of the feature recording data. The assignment system is used to assign programmable motion device handling parameters to marker sensing data based on the association data. The assignment system includes a workflow management system and a separate operable controller. The programmable motion device handling parameters include any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position where the object engages with the end effector, and the position where the end effector holds the object when it is grasped.

[0009] According to another aspect, the present invention provides an object classification system for an object handling system including at least one programmable motion device. The object classification system includes an analysis system, a handling parameter system, and an assignment system. The analysis system includes at least one characteristic sensing system for providing sensing data about the object to be processed. The characteristic sensing data includes data related to any one of weight, height, width, length, center of mass, object description, object category, and at least one image. The handling parameter input system is used to obtain handling parameter input data about the object. The handling parameter data includes data related to any one of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid. The assignment system is used to assign programmable motion device handling parameters to the tag sensing data. The assignment system includes a workflow management system and a separate operable controller. The programmable motion device handling parameters include any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position where the object engages with the end effector, and the position where the end effector holds the object when it is grasped.

[0010] According to another aspect, the present invention provides an object classification system for an object processing system including at least one programmable motion device. The object classification system includes a tag sensing system, an analysis system, a transport parameter input system, a non-transitory storage medium, and an assignment system. The tag sensing system provides tag sensing data regarding identification tags of the object to be processed. The analysis system includes at least one characteristic sensing system for providing sensing data regarding the object to be processed. The characteristic sensing data includes data related to any one of weight, height, width, length, center of mass, object description, object category, and at least one image. The transport parameter input system obtains transport parameter input data regarding the object. The transport parameter data, including the data, is related to any one of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid. The non-transitory medium stores the characteristic sensing data and transport parameter input data associated with the tag sensing data. The assignment system assigns transport parameters of the programmable motion device to the tag sensing data. The programmable motion device handling parameters include any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position where the object engages with the end effector, and the position where the end effector holds the object when it is grasped. Attached Figure Description

[0011] The following description can be further understood with reference to the accompanying drawings, in which:

[0012] Figure 1 An illustrative diagram view of an inductive system including an input system according to one aspect of the present invention is shown;

[0013] Figure 2 It shows Figure 1 An illustrative diagram of the weight sensing conveyor section of the input system;

[0014] Figure 3 It shows Figure 2 An illustrative side view of the weight-sensing conveyor section;

[0015] Figure 4 It shows an unknown object on it. Figure 2 An illustrative diagram of the weight-sensing conveyor section;

[0016] Figure 5 It shows Figure 4 Illustrative side view of the weight-sensing conveyor section and the object;

[0017] Figure 6 It shows Figure 4 Illustrative plan view of the weight-sensing conveyor section and the object;

[0018] Figure 7 It shows Figure 4 Illustrative end view of the weight-sensing conveyor section and the object;

[0019] Figure 8 An illustrative schematic view of a weighing system with multiple rollers is shown in a system used for one aspect of the present invention;

[0020] Figure 9 It shows Figure 8 An illustrative side view of the scale system;

[0021] Figure 10 It shows Figure 8 An illustrative diagram of a part of the weighing system, open to the view;

[0022] Figure 11 An illustrative diagrammatic view of a response evaluation section in a system according to one aspect of the present invention is shown;

[0023] Figures 12A to 12D This shows what happens when an object moves on the response evaluation section. Figure 11 An illustrative side view of the response assessment section;

[0024] Figure 13 An illustrative diagram illustrating the operation of a system according to one aspect of the present invention is shown;

[0025] Figure 14 An illustrative diagrammatic view showing an example of the characteristics of the qualification or transport parameters in a system according to one aspect of the invention;

[0026] Figure 15 An illustrative diagrammatic view of an inductive object processing system according to one aspect of the present invention is shown;

[0027] Figures 16A to 16C An illustrative schematic view of a bin according to one aspect of the invention is shown, showing a camera view ( Figure 16A ), volume scan of the hopper ( Figure 16B ) and 3D scanning using edge detection ( Figure 16C );

[0028] Figure 17 An illustrative diagram view of an inductive system including a closed scanning system according to another aspect of the present invention is shown;

[0029] Figure 18 It shows Figure 17An illustrative side view of a closed scanning system (part of the housing has been removed for clarity);

[0030] Figure 19 It shows Figure 17 Illustrative side view of a multi-frequency inspection station in a closed scanning system;

[0031] Figure 20 It shows Figure 19 A schematic bottom view of a multi-frequency inspection station;

[0032] Figure 21A and Figure 21B It shows Figure 19 A side sectional view of a low-pressure X-ray scanning system in a closed scanning system;

[0033] Figure 22 It shows a probe with a negative charge. Figure 19 An illustrative side view of a triboelectric scanning system for a closed scanning system;

[0034] Figure 23 It shows a probe with a positive charge. Figure 19 An illustrative side view of a triboelectric scanning system for a closed scanning system;

[0035] Figure 24 It shows Figure 19 An illustrative plan view of probe movement in a triboelectric scanning system;

[0036] Figures 25A to 25H An illustrative diagram view of an SKU deduction processing system according to one aspect of the present invention is shown;

[0037] Figure 26 An illustrative diagrammatic view of a functional control system in an inductive system according to one aspect of the present invention is shown;

[0038] Figure 27 An illustrative graphical view of a parameter estimation application for a system according to one aspect of the invention is shown; and

[0039] Figure 28 An illustrative schematic diagram of a processing system with an inductive system according to an aspect of the present invention is shown, the inductive system comprising multiple processing stations.

[0040] The accompanying drawings are shown for illustrative purposes only. Detailed Implementation

[0041] According to some aspects, the present invention provides a system and process for grouping objects into an automated object handling system. For example, a warehouse worker receives previously unprocessed SKUs. The worker scans a barcode to identify the object. The worker then uses a machine to measure the object's height, width, length, and weight (such as one sold by Quantronix, Farmington, Utah). A scanner scans the object. Personnel can then add additional information about the SKU (such as whether it is a fragile item), which can help with SKU classification or determining handling parameters. This additional information can be entered in the GUI or via a form that displays images and barcodes (discussed below). Personnel then empty the box containing the new SKU into the crate, and because the entire process is linked to both the SKU barcode and the crate barcode, the database now contains information about the crate's contents.

[0042] In the background, weight and size data, along with additional information provided by personnel, are used on a separate server to calculate whether an SKU qualifies for automation, and what the handling parameters are (e.g., optionally involving machine learning to extrapolate on new SKUs). These parameters are then used to feed the SKU to either a manual picking unit or a specific automated picking unit.

[0043] Current object handling warehouses have existing equipment and processes built entirely around manual processes that have been in use for decades. While some automated systems force warehouses to adapt to new automated equipment (e.g., adding lines to the floor for robots to move around, using X scanners, placing items at Y height, etc.), the systems of various aspects of this invention provide equipment adapted to existing warehouses.

[0044] When adding a robotic picking system to an existing warehouse, a suitable system is needed to prepare items for robotic picking. Some warehouses receive truckloads of sealed cardboard boxes and handle everything inside entirely manually (e.g., people move the boxes, cut them open, and place individual items into new boxes). In these cases, robots cannot directly replace humans. Further process optimizations are required to ensure the efficiency and functionality of the robots. The tasks typically include: First, selecting objects eligible for robotic picking. Some items may be too heavy, too large, too small, too light, etc., while others may be improperly packaged. Sometimes eligibility can be determined using item data, but other times, visual inspection of objects received from suppliers is necessary. Second, the objects must then be placed in optimal containers for delivery and presentation to the robot for robotic picking. Robots cannot pick objects from any container. For example, items tightly packed in cardboard boxes may be difficult to pick by a robot.

[0045] The systems disclosed in this paper illustrate ways to handle these two tasks (as well as other tasks that can be considered as added value in warehousing and / or robotic picking). Many of the disclosed ideas combine automated and manual processes, thus using a balance to capture the advantages of each. The result is a net reduction in labor while increasing system efficiency. The goal is to increase throughput.

[0046] SKU is a stock unit and represents one item. Each type and size of shampoo represents a different SKU. An object (e.g., a unit or per) is one of many things. A box is typically a cardboard box that holds multiple units of the same SKU. Quality Assurance (QA) typically refers to the part of a retail organization whose task is to ensure that the right goods are delivered intact to the right place. A Warehouse Management System (WMS) is a computer system that communicates with robotic units. The WMS tells the robots what orders to fulfill, for example, “pick an item from the bin with barcode Z and put it in the empty container, then indicate what its barcode is.” A Robot Control System (RCS) centralizes information about SKUs and communicates to all robots operating at the site what to do when an SKU (new or previously encountered) is received. An Automated Storage and Retrieval System (AS / RS) is an automated system for storing and retrieving bins; in a warehouse environment for store replenishment or e-commerce, these bins hold one SKU or are subdivided into compartments, each holding one SKU.

[0047] There are two important parameters: automation qualification and handling parameters. Automation qualification involves flags indicating whether a given SKU will be sent to a robot (or, if multiple different types of robots are available, to one type). Handling parameters involve what parameters will be used when handling the SKU (if the object is qualified for automation), such as robot speed, suction cup or vacuum pressure, and any other parameters. Both parameters need to be set for all SKUs that the robot cell may encounter. Warehouses typically have hundreds or thousands of SKUs, so manually setting these parameters is often tricky.

[0048] Systems and processes are provided for determining SKU automation eligibility through execution unit evaluation, depending on various aspects. In one approach, the warehouse management system (WMS) assumes all new SKUs are eligible for automation and can be fed to the robot. The robot then attempts to pick all SKUs and detects poor performance in picking a particular SKU based on: (a) dropping; (b) incorrect picking; (c) multiple picking; (d) inability to empty the bin; or (e) any other metric indicating abnormal performance. In this case, the SKU or the bin containing the SKU can be tagged and automatically sent to QA (building QA, dumping station, or other location) for a manual automation eligibility decision.

[0049] The timing of these evaluation picks can be optimized or scheduled. There is a risk of wasting time on initial evaluation picks, as the robotic system may pick incorrectly, drop items, etc. On one hand, these new SKUs can be evaluated during periods of lower system downtime throughout the day. These new SKUs can be evaluated while simultaneously weighing costs (the possibility of errors at the unit and manual intervention) against benefits (expanding the number of SKUs that can be picked automatically), while also meeting operational metrics (meeting daily throughput targets). Furthermore, these new SKUs can be based on metrics such as how often a given item might be picked, based on information about SKU speed (i.e., order frequency).

[0050] When in this new SKU evaluation mode, the system may operate more conservatively than with previously encountered SKUs. The robot may move more slowly or change parameters in other ways, such as taking longer. The system may also perform exploratory movements, such as gently shaking the item without actually performing a full transfer movement, to further evaluate the SKU's handling capabilities. During these evaluation picks, the unit can notify personnel or operators that it is in evaluation mode, allowing them to retrieve the item more quickly if needed.

[0051] On the other hand, systems and methods involve using customer return data or other customer feedback (good or bad). Such systems and methods can use customer return information to inform handling parameters or automation eligibility. In this case, the WMS sends categorical or qualitative information from the receiving customer or retail store to the RCS, such as: satisfaction with received items, items being damaged, or incorrect quantity of items received. This information can be associated with a specific SKU, or, in the case of e-commerce, a specific customer order, or a box / carton of items sent to the store. Once the exact unit can be identified, all sensor data and handling parameters for that specific unit can be correlated with other objects and used as input to machine learning algorithms that better select handling parameters to avoid damage or generally improve customer satisfaction.

[0052] According to another approach, systems and methods are provided for training personnel and robots using picking and dumping test stations. In this case, instead of assuming all new SKUs are eligible for automation, new SKUs are checked through a separate automation system. For example, the system could repeatedly and automatically pick items to test their automation eligibility. This would be a test unit not part of the usual flow of goods throughout the warehouse. The operation could be bin-to-bin, i.e., filling the left bin with the SKU and transferring all units to the right bin. Once the left bin is empty, the items are transferred back to the right bin; this is repeated multiple times to determine the compatibility of the new SKU and other picking parameters. The cost of errors at this unit is much lower and does not affect operations. Alternatively, this could be a semi-automated or fully manual process. A manual gripper test station (equivalent to a hose with suction cups) can be used to attempt picking, which would evaluate whether the system can grasp items. In this system, personnel can also train robots on how to grasp items. The test station can be used by personnel to determine dumping decisions (preferably face up), including, for example, other dumping instructions (maximum number of bins per bin, etc.), and whether objects can be handled without the test station.

[0053] For example, Figure 1 A SKU summarization system 10 is shown, which processes and records data about objects fed into an object processing system. The summarization system 10 includes an input station 14 to which new objects 11, 13, 15 are presented, for example, in a single flow on a conveyor 40. For example, objects may be provided in cases of similar new objects. One object from each case may be presented for entry into the system. Any conveyor in the system discussed herein can be a patterned or unpatterned conveyor, and the system may monitor the movement of the conveyor (and thus the objects on it) via multiple sensors and / or a conveyor speed control system. The summarization system includes a weight-sensing conveyor system 12, a response evaluation section 16, an information input system 18, and a bin preparation system 20, for example, which includes a programmable motion device 19 for placing objects into bins 21 for processing by the object processing system.

[0054] The weight-sensing conveyor system includes a conveyor section 22 mounted on rollers 24 and 26, with a weighing sensor 28 and 30 mounted at both ends of each roller (only one weighing sensor is shown at one end of each roller 24 and 26). Figure 2 Side view and Figure 3An isometric view is shown. Damaged packaging can also be identified by the sensing system, for example, if the packaging appears wet or leaking. By incorporating humidity sensors into the weighing sensors 28, 30, humidity-sensitive sensors can be used in conjunction with conveyor 40 in any pretreatment system discussed herein. In other embodiments, a camera capable of detecting humidity (e.g., a trillion-fps camera capable of tracking photons) can also be used in such an inductive system. Detection of any dampness indicates that the object may be damaged, thus requiring exceptional handling.

[0055] refer to Figure 4 According to another aspect of the invention, the system can further provide a processing system having a conveyor section 22 (where the object 32 is located on the conveyor section 22) that can not only determine the weight of the object 32, but also (further reference) Figures 5 to 7 The difference between the length end and the width end, as well as the weight sensed by each weighing sensor 28 to 31, can be used to determine the center of mass region of the object 32. Figure 5 A side view of an object on a conveyor along the length L of a conveyor section is shown, and Figure 7 An end view of an object on a conveyor along the width W of a section of the conveyor is shown. Figure 6 A plan view of the conveyor section is shown.

[0056] Figure 8 A weighing system 40 is shown, which includes a plurality of rollers 42 within a frame 44 mounted on a base 46. Further reference... Figure 9 Side view and Figure 10 An enlarged view shows each roller 42 mounted to the frame 44 via a load cell or torque sensor 48 at either end of each roller 42. System 40 can be used in any system discussed herein. By monitoring the output of each of the load cells or torque sensors 48, the center of mass of the object on the roller can be determined.

[0057] Therefore, such systems providing weight sensing in a presentation conveyor may include one or more load cells or weight-sensitive mechanisms embedded in a surface on which an object is presented to a programmable motion device, such as an articulated arm. The weight and / or observed density (weight / volume) of each object can be estimated using a camera or distance sensor of a volume-sensing programmable motion system.

[0058] Figure 1 The response evaluation section 16 includes one or more sets of transport rollers 50 and one or more disturbance rollers 52, such as Figure 11 As shown. Further references Figures 12A to 12DSensing unit 54 (e.g., camera or scanner) is horizontally pointed at conveyor section 16, and sensing unit 56 (e.g., camera or scanner) is downward pointed at conveyor section 16. When object 58 travels along conveyor roller 30 (e.g....), Figure 12A As shown), it will contact the disturbance roller 52 (as shown). Figure 12B (As shown). The disturbance roller 52 can be any roller with a large diameter, or it can be raised relative to the transport roller 50 and rotate at a faster speed than the transport roller 50. Figure 12C An object on the disturbance roller 52 is shown, and Figure 12D An object has been discharged from the agitation roller 52. Observing the object's responsive movement as it travels onto and off the agitation roller provides information about the object's weight, composition, and balance. In this way, and using sensing units 54, 56, the system can determine (in conjunction with the computer processing system 100) various characteristics of the object 58. For example, roller 52 can be mounted on a torque sensor (as discussed further above), and when it is determined (using sensing unit 54) that the object 58 is balanced on roller 52, the system can determine an estimated weight. Therefore, roller 52 on the torque sensor can be used to determine the weight of the object as it passes over the roller.

[0059] Furthermore, if roller 52 rotates at a faster rotational speed, the system can determine the inertia value of object 58 as it engages and disengages from the roller. Various other characteristics, such as the center of mass (COM), can also be determined or estimated using the roller in conjunction with the sensing unit, as discussed further herein and below. The system can also use the sensing unit and roller 52 (along with the computer processing system 100) in the same way to determine whether an object is a collapsible bag and / or whether the assumed object 58 is actually a multiple-pick (including more than one object) by observing whether the object moves separately and / or whether the shape of the object changes as it crosses roller 52. According to other aspects of the invention, the transport roller 50 can be replaced by a conveyor section located below the height of the disturbance roller 52.

[0060] On the other hand, systems and methods are provided for generating handling parameters that provide information in a semi-automatic manner. In this case, personnel inspect the SKU, perform measurements such as recovery weight and product dimensions, and then provide classification information about the SKU. A software system then receives all this information and determines automation eligibility and handling parameters (using regression, machine learning, or other algorithms). Personnel scan item barcodes to obtain the SKU and then put it into a weighing and dimensioning system (e.g., as described above). (Scanner). Then, the staff will categorize all relevant information about the SKU as characteristics, such as: fragile, brittle, or easily damaged (this could mean how forcefully you move or push the item down), in a plastic bag (this could mean how fast you move it), rolled up (this could mean it might roll when placed), not stackable (this could mean don't put it at the bottom of a pile of things), glass (so don't handle it with a robot, or handle it very gently), very thin (so don't handle it if it's incompatible, or use a special type of gripper), and opens when picked up (like a book or jeans, and therefore shouldn't be handled by a robot).

[0061] For example, Figure 13 An operational diagram of a system according to one aspect of the invention is shown. As shown at 60, the process begins (shown at 62) by receiving a new box containing the SKU, and then (shown at 64) by scanning the SKU of the new box. The system then (shown at 66) measures the weight and dimensions, as discussed above (or using the methods mentioned above). (Scanner). The weight and size information is then provided to the local WMS 68 and a separate operable controller 70. This is then handled by personnel (e.g., in...). Figure 1 The information input system (at point 18) inputs (shown at point 72) the classification information and provides it to the local WMS68. Then the box (shown at point 74) is emptied into the inventory container.

[0062] A separate operable controller 70 can then send transport instructions 76 and object handling parameter information to any of the various programmable motion devices 90, 92, 94 (as well as data indicating that the object is not eligible for automated handling 96). The system can also periodically mark previously registered SKUs that need to be reweighed and their dimensions analyzed for quality control purposes, as shown at 98.

[0063] Figure 14 Examples of features that can be used to notify eligibility or handling parameters, such as placement permissions or orientation permissions, are shown. Placement permissions are the ability to place an item in a desired location and orientation. Some items may not be easily oriented, or the final placement may be unpredictable. For example, an item loosely packed in a bag may wrinkle or fold on its own, or it may swing / wobble uncontrollably before placement, making its final size and orientation unpredictable. Orientation permissions are the ability of an object to maintain its placement position and orientation. Information can be entered by scanning the scan form 100 (e.g., via...). Figure 1 Information input station 18 (digital input). In particular, after scanning the SKU, the code on the form can be scanned (e.g., using scanner 110) to provide information about the object. Figure 14At position 102, code that can be scanned (after scanning the SKU) to indicate that an item is rollable (with low placement permissions) is shown, and at position 104, code that can be scanned to indicate that an item is not stackable is shown. Figure 14 At position 106, a code that can be scanned to indicate that an item is fragile or crushable is shown, and at position 108, a code that can be scanned to indicate that an item is deformable (with low placement permissions) is shown.

[0064] Figure 14 At 112, a code is shown that can be scanned (after scanning the SKU) to indicate that an item is too thin or too small to be reliably held, and at 114, a code is shown that can be scanned to indicate that an item is glass or other fragile. Figure 14 At 116, code that can be scanned to indicate that an item is open (with low placement permissions) is shown, and at 118, code that can be scanned to indicate that an item is too heavy to be processed by an automated processing system is shown.

[0065] End effectors can be used with programmable motion devices in object handling systems. For example, Figure 15 An object handling system 120 is shown, comprising an object handling station 122 located between a feed conveyor 124 carrying a feed bin 126 and a destination conveyor 128 carrying a destination container 130. The object handling station 122 includes a programmable motion device (e.g., an articulated arm 132) having an attached end effector and an associated sensing system 134. The sensing system 134 is positioned to sense objects (and / or associated markers) in selected feed bins 126', which are turned (selected) by the bidirectional conveyor 136 to move onto selected feed conveyor sections 124'. The sensing system 134 is also positioned to sense destination containers 130' disposed on the destination conveyor sections 128' of the destination conveyor 128' via one or more deflectors, which selectively turn selected destination containers 130' onto the destination conveyor sections 128'. The operation of the system is controlled by one or more computer processing systems 200, which communicate with the conveyors 124, 124', 136, 128, 128' and the programmable motion device 132 (including an end effector) and the sensing system 134.

[0066] The object handling station 122 includes a feed conveyor section 124' that uses a bidirectional conveyor 136 to circulate selected supply bins 126' back and forth. The end effector of the programmable motion device 132 is programmed to grab objects from the supply bins 126' and move them to deliver them to the desired destination bin 130' on the destination conveyor loading area 128' by placing or dropping the objects into destination containers 130' on the destination conveyor loading area. The supply bins 126' can then be returned to the input conveyor 124 and optionally taken to another handling station. Thus, at the handling station 122, one or more supplier supply bins 126' are conveyed to the input area, and the programmable motion device 132 is actuated to grab objects from the bins 126' and place them into the selected destination containers 130'. The processed supplier container 126' is then returned to the common input stream on the conveyor 124, and the destination container 126' moves further along the destination conveyor 124.

[0067] System 120 may also include one or more sensing units 138 located on or near the feed conveyor for identifying markings on the exterior of each bin, providing sensing data from which the contents of the bin can be identified, and then determining its relative position on the conveyor 124 to track its location. According to one aspect, it is assumed that the bins of objects are marked with visually unique markings, such as barcodes (e.g., providing UPC codes), QR codes, or radio frequency identification (RFID) tags or mailing tags, making them readily identifiable by a scanner for processing. The type of marking depends on the type of scanning system used but may include 1D or 2D code symbols. Various symbol methods or tagging methods can be employed. It is assumed that the type of scanner used is compatible with the marking method. For example, marking by barcodes, RFID tags, mailing tags, or other means encodes the identification mark (e.g., a string of symbols), which is typically a string of letters and / or numbers. The string of symbols uniquely associates a supplier bin with a specific set of similar objects. Based on the identification code on the feed bin 124, the system can allow the bin 124 to continue along the feed conveyor 124, or can guide the selected bin 126' onto the selected feed conveyor 124'.

[0068] At the selected feed conveyor 124' at the object handling station 122, the sensing system 134 assists (using the central control system 200, such as one or more computer processing systems) the programmable motion device 132, including an end effector, to locate and grasp objects in the feed bin 126'. Depending on other aspects, each object may also be marked with a visually unique identifier, such as a barcode (e.g., providing a UPC code), QR code, or radio frequency identification (RFID) tag or mailing label, making them readily identifiable by a scanner for handling. The type of marking depends on the type of scanning system used but may include 1D or 2D code symbols. Similarly, multiple symbol methods or tagging methods may be employed on each object.

[0069] System 120 also includes a field-capture sensing system 140, which includes multiple sensing units 142, 144, 146, 148 pointing downwards at one or more objects in each feed bin 126 on the feed conveyor 124, and a weight sensing section 139 of the conveyor 124 below the sensing system. Furthermore, the weight sensing section 139 may also include a vibration device 137 for shaking the bin to separate objects within the bin, as discussed in more detail below. The sensing system is mounted above the conveyor, facing each bin containing the next object to be transported, thus looking down at each bin 126. Sensing units may include, for example, cameras, depth sensors, and lights. A combination of 2D and 3D (depth) data is acquired. The depth sensor provides depth information, which can be used in conjunction with camera image data to determine depth information about various objects in the view. Lights can be used to remove shadows and facilitate the identification of object edges, and may be turned on all at once during use, or may be illuminated in a desired sequence to aid object identification. The system uses this image and various algorithms to generate a set of candidate gripping locations for objects in the bin, as discussed in more detail below.

[0070] Figure 16A A view of a bin 126 from a sensing system 140 is shown, which includes sensing units 142, 144, 146, and 148. Figure 16A The image view shows bin 126 (e.g., on conveyor 124), and bin 126 contains objects 150, 152, 154, 156, and 158. While in some systems the objects in each feed bin may be dissimilar (multiple SKUs), in other systems, such as... Figure 16AAs shown, the objects may be of the same type (single SKU). The system will identify candidate gripping locations on one or more objects and may not attempt to identify gripping locations on objects partially occluded by other objects. A 3D model of a robotic end effector, positioned where the actual end effector will be used as the gripping location, can be used to indicate candidate gripping locations. For example, a gripping location may be considered good if it is close to the object's center of mass to provide greater stability during gripping and transport, and / or if it avoids locations on the object where a good vacuum seal may not be possible (such as lids, seams, etc.).

[0071] The sensing system 140 includes scanning and receiving units and edge detection units in sensing units 142 to 148 for capturing various characteristics of selected objects throughout the bin. Similarly, Figure 16A The diagram shows a view from a capture system, which, according to one embodiment, may include a group of similar or identical objects 150, 152, 154, 156, 158. Scanned volume differences (if any) are shown in... Figure 16B The data is shown in the diagram and compared with recorded data on items identified by identification tags provided by the detection system of the SKU induction system, or recorded object data. In particular, the scanned volume is compared with the volume of the identified SKU multiplied by the known number of objects in the bin.

[0072] Depending on other aspects, the scanning and receiving unit can also be used to determine the density of the object set in the bin, comparing this density with a known density of the identified SKU multiplied by a known number of objects in the bin to determine the mass and volume of the objects. Volume data can be obtained, for example, using a light detection and ranging (LiDAR) scanner, a pulse time-of-flight camera, a continuous-wave time-of-flight camera, a structured light camera, or a passive stereo camera.

[0073] Depending on other aspects, the system may additionally use an edge detection sensor (also in conjunction with processing system 200) to detect the edges of any objects in the bin, for example using data on intensity, shadow detection, or echo detection, and may be used to determine, for example, the edges of objects in the bin. Figure 16C The size, shape, and / or outline shown may help confirm the quantity of objects in the bin. In some respects, the system can identify specific objects in the bin and confirm their shape and size through such edge detection. Therefore, the system described above can be used to confirm the quantity of objects in the bin, and in some respects, to initially estimate the quantity of objects (for a single SKU) in the bin and / or to confirm recorded data for any particular SKU.

[0074] Similarly, the operation of the aforementioned system is coordinated with a central control system 200, which also communicates (e.g., wirelessly) with the articulated arm 132, sensing systems 134, 138, and 140, as well as the feed conveyors 124, 124', bidirectional conveyor 136, destination conveyors 128, 128', and any steering mechanisms. The system determines the UPC associated with the supplier's bin and the outbound destination for each object based on a symbolic string. The central control system 200 consists of one or more workstations or central processing units (CPUs). For example, the mapping between UPCs or mailing tags and outbound destinations is maintained by the central control system in a database called a manifest. The central control system maintains the manifest by communicating with a warehouse management system (WMS). The manifest provides an outbound destination for each inbound object.

[0075] Figure 17 A SKU deduction system 150 according to another aspect of the present invention is shown, the deduction system comprising as referenced above. Figure 1 The input station 14 discussed presents new objects 11, 13, 15, for example, in a single stream on conveyor 40. Objects can be supplied from bins of similar or dissimilar objects, with one object presented at a time for aggregation into the processing system. Similarly, any conveyor can be patterned or unpatterned, and the system can monitor and control the movement of the conveyor via speed control. Aggregation system 150 includes the weight-sensing conveyor section 12 discussed above, the response evaluation system 16 discussed above, and bin preparation system 20, all of which, for example, include programmable motion devices 19 for placing objects into bins 21 for processing by the object processing system. Aggregation system 150 also includes a closed-loop scanning system 152 through which conveyor 154 transports objects en route to bin preparation system 20.

[0076] refer to Figure 18 The enclosed scanning system 152 may include multiple scanning systems 160, 162, and 164, for example, scanning an object as it passes through the system 152. The scanning systems 160, 162, and 164 may be timed to operate at different times and may be separated by isolators 166 and 168, and in some respects may be further enclosed in a separate housing.

[0077] Further reference Figure 19 and Figure 20System 160 may include multiple light sources 170 (e.g., laser sources of different frequencies). Illumination reflected from the (highly reflective) surface of object 174 can be immediately returned to a nearby sensor 172, while illumination entering and returning from object 174 can be returned to a sensor 176 further radially. System 160 can learn the response of different types of plastics (e.g., bags) at different frequencies to determine various properties of the object material, including, for example, reflectivity and refractive index at different wavelengths. Additionally, the system employs not only light sources of various wavelengths but also infrared or ultraviolet radiation.

[0078] Figure 21A and Figure 21B A scanning system 162 including a portion of an X-ray source is shown. This X-ray source may include an X-ray tube 180 within a housing 182, such that the X-ray tube emits X-rays through an X-ray output region 184 of the housing 182. The X-ray tube includes an anode 186, a cathode 188, and an intermediate section 190 between the anode and cathode 188. The anode 186 of the X-ray tube 10 includes an anode cover 192, an X-ray generating target 194, and an X-ray transmission window 196. The cathode 188 includes a cathode shroud 198, an electron emitter 204, and electrical connections 206 and 208 through which heater power is applied to the electron emitter 204. The intermediate section 190 may be formed of an electrical insulator such as ceramic or glass. The electrical insulator is sealed to the anode and cathode of the X-ray tube, thereby creating an internal region of the X-ray tube in which a vacuum can be generated and maintained.

[0079] The scanning system 162 is positioned above the detector 200, and the object 202 (such as a portion of a shipping bag, e.g., an edge or corner) can be positioned above the detector 200. When the scanning system 162 is near the object 202, the X-ray scanner is activated, and heater power is supplied to the cathode electron emitter 204. This applies a high voltage (e.g., 30 to 50 kV) between the cathode end 188 and the anode end 186. The electric field generated by the applied high voltage accelerates electrons from the electron emitter through a vacuum to the X-ray generating target 194. The intensity of the X-rays generated at the target increases with increasing high voltage, electron beam current, and atomic weight of the target material. A portion of the X-rays generated in the target exits the tube via the X-ray transmission window 196 and exits the housing 182 via the X-ray output region 184 of the housing 182. The high voltage at the cathode end is typically provided as a negative high voltage (e.g., -50 kV), and the voltage potential at the anode end is typically provided at the system's reference ground potential. This allows the anode 186 of tube 180 to be directly coupled to housing 182. X-ray tube 180 can be packaged in a reciprocating device that includes a high-voltage power supply and a power supply for driving the electron emitter.

[0080] Depending on the power level adjustment, the scanning system 162 can be used to determine either the material or density of the shipping bag and / or any contents. For example, in some embodiments, the system 162 can be used to distinguish low-density polyethylene (0.917 to 0.930 g / cm³). 3 ) and high-density polyethylene (0.944 to 0.065 g / cm³) 3 This system can also be used to determine whether an object's density is too high for the articulated arm to safely lift or move it, or in other ways to determine comparative responses to a variety of known materials for machine learning purposes.

[0081] exist Figures 22 to 24 The image shows a scanning system 164, which includes a triboelectric scanning system comprising an electrically charged triboelectric probe 220 attached to a movable shaft 222. (See image for details.) Figure 22 As shown, probe 220 can be negatively charged, and as Figure 23 As shown, according to different aspects of the invention, probe 220' may be positively charged. System 164 also includes one or more (e.g., as shown) on the elongated conductive rod 226. Figure 24 The four detection probes 224 are shown. When probes 220 and 220' are placed on the surface of object 229, as... Figure 24 As shown in the top view, probes 220 and 220' are moved (e.g., through a cyclical motion to move them closer to and away from various detection probes 224), any detected positive or negative charge (electron loss or electron gain) is detected by one or more detection probes 224. Since different materials behave differently to the presence or absence of additional electrons, material distinctions can be determined. For example, known materials (e.g., plastic materials) become more negatively charged in a list of acrylic, polystyrene, polyurethane, polyethylene, polypropylene, and ethylene resins. Similarly, using machine learning, the responses of different materials to probes 220 and / or 220' can be correlated with later learned or known material responses to facilitate the identification of the material of an object or the shipping packaging of an object.

[0082] refer to Figures 25A to 25H According to one aspect, the process for providing SKU summarization may involve processing an object if an SKU is identified, and such processing may involve identifying whether the SKU requires any special handling by a dedicated programmable motion device, or whether the SKU can be handled by a general-purpose programmable motion device. (Reference) Figure 25A Therefore, the process can begin by determining whether an SKU has been identified (step 1000). If the SKU is unknown, the system may attempt to identify the object's weight, size, and / or shape (step 1002), and refer to... Figure 25BThe system may attempt to determine (e.g., from each view of the object, refer to step 1018) the characteristics of different surfaces and / or views of the object. For example, the system may determine whether any shape or surface from each view includes a rectangular shape (step 1020), and if so, increase the positioning authority (ability to hold in place during placement) factor (step 1022). The system may also determine whether any shape or surface from each view includes a circular or rounded surface (step 1024), and if so, decrease the positioning authority factor (step 1026). Furthermore, the system may determine whether any shape or surface from each view includes a shape that is significantly larger than other shapes (step 1028), and if so, increase the factor related to whether the object is too thin or too fine to handle (step 1030). Additionally, the system may determine whether any shape or surface of the object changes significantly when lifted (step 1032), and if so, decrease the placement authority (ability to hold in place during placement) factor (step 1034). Further reference Figure 25C The system can then use the above information to estimate the shape of the object by associating it with other known SKUs (step 1036), determine the weight of the object (step 1038) and center of mass (step 1040), and determine whether the center of mass of the object is not constant (step 1042), for example, whether the object is unstable.

[0083] The system can then return to Figure 25A The text / image / bag analysis routine (step 1004) is used, and then the object packaging is analyzed as follows and referenced. Figure 25D The system can determine the outermost point of the object (step 1044), and then determine which points form right angles or acute angles (step 1046). The distances between such points are then determined (step 1048), and the perimeter is estimated (step 1052), which is used to estimate the shape of the object (step 1052). The length profile along the conveyor direction is then evaluated (step 1054), the width profile is evaluated (step 1056), and the height profile is then evaluated (step 1058), and this information can be used to estimate the volume of the object (step 1060).

[0084] In addition, and refer to Figure 25EFor each view (step 1062), the system employs text and image recognition (step 1064) to detect any words “fragile,” “breakable,” “glass,” “hazardous,” “dangerous,” “flammable,” and “this end up” (step 1066). If any of these terms are detected, the system identifies the object as unprocessable by the programmable motion device (step 1068), and the object is conveyed to the exception bin. The system can also detect any image from various images that may indicate the object is fragile or contains any of the hazardous substances, such as… Figure 25E As shown. If any of these are found on an object, the object is marked as unsuitable for processing by a programmable motion device.

[0085] Additionally, and refer to Figure 25F The system can use a combination of image and text recognition (step 1070) to first identify the plastic recycling symbol and the numbers within it. Since the numbers identify the type of plastic, the system will perform the following associations (step 1072): if the number is 1, the material = polyethylene terephthalate (PET); if the number is 2, the material = high-density polyethylene (HDPE); if the number is 3, the material = polyvinyl chloride (PVC); if the number is 4, the material = low-density polyethylene (LDPE); if the number is 5, the material = polypropylene (PP); if the number is 6, the material = polystyrene (PS); and if the number is 7, the material remains unknown. The system will also search for any terms: “LDPE”, “LD-PE”, “PE-LD”, “HDPE”, “HD-PE”, “PE-HD”, “PET”, “PVC”, “PP”, and “PS” (step 1074).

[0086] The system can then (and refer to) Figures 17 to 24 , Figure 25G and Figure 25H The edges of the object are determined (step 1076) and selected for analysis (step 1078). The object can then be queried using an optical detection head (step 1080), as described above, for example, using infrared, visible, or ultraviolet radiation (step 1082). Various analyses can be performed, including light scattering and refractive index analysis (step 1084).

[0087] The system can then employ edge X-ray analysis (step 1086), X-ray scattering (step 1088), and apply X-ray detection analysis (step 1090) to characterize any scattering or transmission detection to further characterize the object. The system can also (and refer to...) Figure 25HThe material is brought into contact with an electrode (step 1092), and then electrostatic induction analysis (step 1094) is used to determine the triboelectric response to the application (or absence) of electrons on the surface of the material (step 1096). The system can then estimate any of the following: the density of the object (step 1098), viscosity (step 1100), molecular weight (step 1102), and material composition (step 1104). See again... Figure 25A The system can then associate the object with similar or most similar SKUs based on a ranking factor (step 1006) and determine if the similarity ranking with the next closest object is high enough to process the new SKU (step 1008). If the similarity ranking is too low, the system redirects the object to the exception bin (step 1012). If the object needs to be processed, the system determines whether any special handling is required (step 1010). If yes, the object is handled using special handling (step 1014), and if not, the object is handled using a general-purpose programmable motion device (step 1016).

[0088] According to other aspects, the present invention provides strategies for dumping by robots, automated systems, or personnel to improve picking performance. Dumping is the operation of transferring units arriving in a bin to a crate; the crate is then directly transferred to one or more robotic units, or storage units for an automated storage and retrieval system (AS / RS). The performance of the robotic unit can depend on the organization of the objects within the crate. If the objects are neatly laid out and adjacent to each other, the grasp detection algorithm may struggle to generate a pick that selects only one item. For example, if the system captures an image of the crate from above and then selects a circled area (where the suction cup is located), the circle is likely to contain two objects (leading to multiple picks). Accordingly, the robot or automated dumping station, or station for improving picking capacity, may, for example, shake the crate to reduce the effects of tiling, and, for example, spread out the objects in the crate to help the robot make better grasping decisions. Depending on various other aspects, personnel may select certain gripping points on the image to help train the robot, for example, as disclosed in U.S. Patent No. 10,625,432, the disclosure of which is incorporated herein by reference in its entirety.

[0089] Depending on other factors, similar options may be available. Figure 1The SKU sorting system 10 comprises multiple specialized sorting systems for SKU tuning. Different sorting systems are dedicated to handling different types of SKUs. For example, SKUs requiring additional work (e.g., due to their size, low placement authority, or low placement authority). Other SKUs are dumped at faster SKU sorting systems. In other aspects, to provide personnel with information that can aid in this differentiation, one aspect of the system can record objects and successful and unsuccessful dumping frames via photographs. According to another aspect, the system can provide the dumper with a heatmap of successful picks to help the system make orientation decisions and allow personnel to demonstrate gripping points that the robot can later use. The system can also provide instructions on optimal dumping or picking by the robot, which may depend on the SKU, presented to the user on a screen. According to another aspect, the system can use a camera to detect whether objects are being handled correctly and / or project instructions into the workspace, for example, providing a cutting position projected onto the box to be dumped. According to another aspect, personnel are evaluated based on their speed and the number of times the robot fails to pick on their frames, thus motivating personnel to dump in a way that is faster for the robot, not just for the person.

[0090] Depending on other factors, the system can provide suboptimal picking performance or characteristics for SKUs that do not match information in the database, and automatically tag the SKU or basket for delivery to the QA station. This can be detected by the operable controller 70, which can then request personnel to use the system disclosed above or the aforementioned station communication methods. This is done when the scanner weighs or measures an object (SKU). Depending on other aspects, personnel may categorize the menu by packaging type or other items used in compatibility, such as using a crush factor, which is a measure of the susceptibility of an object to crushing.

[0091] Depending on other aspects, the system may evaluate bin characteristics that will improve picking performance. For example, the system may record bin color or texture, such as making the bottom of the bin have a known texture or background. The bin color (e.g., the color of the background or inner surface) can be used for identification purposes, such as sorting bins, or can provide a known background against which objects are more easily identified. Furthermore, a textured inner surface (e.g., wavy) can provide an uneven bottom to aid picking, and / or may include shapes that pull items away from the edges (e.g., inserts), thus again providing a more bowl-like shape to aid picking. Depending on other aspects, the system may provide bins with holes and push blades that can push objects within the bin to aid picking.

[0092] Depending on other factors, the system can provide feedback to inform the supply chain of better packaging options. Some of this information can be used to provide automated hypothesis recommendations based on supplier feedback. Personnel will take photos of examples of substandard packaging to generate reports to send back to suppliers. Additionally, the system can generate scrapes from the photos so partners can see what the scrapes look like.

[0093] Parameter estimation can also be performed throughout the inductive process (e.g., during training and scoring patterns). The scoring pattern uses the current transport parameter estimation model to generate planning parameters for a single SKU or a set of SKUs. Upon receiving new or updated SKU data, the Product Manager will invoke the estimator application.

[0094] The training mode will have the following functions: training of parameter estimation models, offline evaluation of new trained models (see below), comparison of offline performance of new models with production models, decision on whether to generalize the new model to production, and generation of planning parameters for all SKUs in the product collection in the data store.

[0095] The data storage will store a separate experimental test set with a known number of drops, multiple picks, and damages. Within this set, each test SKU will undergo experiments with multiple suction cup sizes and multiple scale durations. This test set will be a subset of the damage and 24 / 7 experiments. The model (either a production model or a newly trained model) will select the optimal parameters from the parameters used for the experiments in the test set for each test SKU. The training application will then count the number of adverse events generated by the selected parameters.

[0096] For each SKU in the product collection, the data store will store two sets of planning parameters: one for inductive and single-pick units, and another for multiple-pick units. The large database will also store two serialization models: one for inductive and single-pick units, and another for multiple-pick units. Training will run on demand and periodically; for example, after new training data is uploaded to the data store, personnel will be able to initiate the training process (both data upload and training initiation will be done through the training application console). Additionally, training can be registered as a scheduled job and, for example, weekly.

[0097] Figure 26A functional control system for use with a system according to one aspect of the invention is shown at 270. This functional control system includes a central control data center 272 that communicates with multiple SKU aggregation stations 274, 276 to provide planning parameters from a data storage 278 (of the central control data center 272) to the separation stations 280, 282. Picking results from the separation stations are provided back to the data storage 278, which also receives manual uploads of 24 / 7 test results and product planning parameters for each type of unit from the product manager 286.

[0098] Data storage 278 also communicates with parameter estimator application 284, and specifically provides, for example, the results of picking and damage experiments to training module 290, which provides the parameter estimation model and offline evaluation results back to data storage 278. Data storage 278 also provides the parameter estimation model to scoring module 292 of parameter estimator application 278, and scoring module 292 communicates with product manager 286 to provide planning parameters for each type of unit and receive product information from product manager 286. Product manager 286 also receives notifications from workflow management system 288 for generating parameters for existing products, as well as new product information and update information for existing products.

[0099] The final production model, used for scoring, provides closed-loop training and lacks access to drop and multiple-pick data, relying solely on SKU features to further predict adverse event scores. This multiple-pick and drop data is referred to as privileged features in this paper. This data is available during training but not during scoring. The table below summarizes the four types of data that will be used in both training and scoring modes.

[0100] Table 1

[0101]

[0102] Dataset A was obtained from dataset C by removing privileged features. The parameter estimation application will be trained using a self-training method. Figure 27 The production model is illustrated in Figure 350, showing the SKU processing steps for dataset scoring and training. Specifically, dataset C (labeled training data with privileged features) 360 is provided to training module 362, which communicates with intermediate module 364 and, in turn, scoring module 368. Scoring module 368 also communicates with dataset D 366, which includes unlabeled training data with privileged features. Scoring module 368 communicates with analysis module 370, which includes the labeled result dataset D 372 and dataset C (labeled training data with privileged features) 374 from dataset C 360.

[0103] Analysis module 370 provides data 376 with privileged features removed to provide dataset A (labeled training data) 378, which communicates with another training module 380. The output of training module 380 is provided to candidate prediction model 382, ​​and candidate prediction model 382 communicates with another scoring module 384, which receives dataset B (unlabeled rating data) 386 and provides SKU parameters 388.

[0104] Figure 28 A system comprising multiple object processing systems 420, 460, 470, 480, and 490 is shown at position 400, as referenced above. Figure 15 The discussed, and any one or two or more such processing systems may include associated field intake sensing systems 440, 450, as also referred to above. Figure 15 The discussed field ingestion sensing system 440. Each field ingestion sensing system 440, 450 can provide inspection of the container as it passes through each system 440, 450, not only visually and volumetrically inspecting the contents, but also providing information on the weight of the contents and redistributing objects within the container for later robot grasping. Each object handling system 420, 460, 470, 480, 490 communicates with a common feed conveyor 504 on which the feed container 502 is provided and a common output conveyor 508 on which the output container 506 is provided.

[0105] Those skilled in the art will understand that various modifications and variations can be made to the above-disclosed embodiments without departing from the spirit and scope of the invention.

Claims

1. An object classification system for assigning transport parameters to objects, the object classification system comprising: An analysis system comprising at least one feature sensing system for providing feature sensing data about an object to be processed with new inventory units, the feature sensing data including data related to any one of weight, height, width, length, centroid, object description, object category, and at least one image; An association system comprising an object information database having characteristic record data, the characteristic record data including data related to any one of weight, height, width, length, centroid, object description, object category, and at least one image of each of a plurality of objects, the association system assigning association data to the object in response to commonality between any of the characteristic perception data and any of the characteristic record data, wherein assigning the association data to the object includes associating the object with similar or most similar inventory units based on a ranking factor, and determining whether the similarity ranking with the next closest object is high enough to process the new inventory unit by a programmable motion device; if the similarity ranking is too low, the object classification system redirects the object to an exception bin; and An assignment system for assigning programmable motion device handling parameters to the new inventory unit of the object based on the associated data, the assignment system including a workflow management system and a separate operable controller, the programmable motion device handling parameters including any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum travel speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position for engaging the object with the end effector, and placement in which the end effector holds the object during gripping.

2. The object classification system of claim 1, wherein a person is permitted to input either additional characteristic perception data or additional characteristic record data into the object information database.

3. The object classification system of any one of claims 1 to 2, wherein the object is configured to have multiple similar objects in a box, and wherein the box includes a box identification mark thereon, the box identification mark being sensed by a box sensing system to generate box sensing data, the box sensing data being provided to the object classification system.

4. The object classification system of claim 1, wherein the object classification system further comprises a handling parameter input system for obtaining handling parameter input data about the object, the handling parameter input data including data related to any of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid.

5. The object classification system of claim 4, wherein either previously recorded characteristic perception data or transport parameter input data of different previously classified objects is accessed by the assignment system to help generate the programmable motion device transport parameters of the objects.

6. The object sorting system of any one of claims 4 to 5, wherein the programmable motion device handling parameters include data regarding transport instructions for transporting the object to a selected programmable motion device among a plurality of programmable motion devices.

7. The object classification system of claim 1, wherein the programmable motion device handling parameters include at least one object processing system having dedicated handling parameters.

8. The object sorting system of claim 7, wherein the dedicated handling parameters include any one of size, weight, placement authority, or arrangement authority.

9. An object classification system for an object processing system including at least one programmable motion device, the object classification system comprising: An analysis system comprising at least one feature sensing system for providing feature sensing data about an object to be processed having a first inventory unit, the feature sensing data including data relating to any one of weight, height, width, length, centroid, object description, object category, and at least one image; A handling parameter input system is used to obtain handling parameter input data about the object, the handling parameter input data including data related to any of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid. An association system includes an object information database with characteristic record data. The association system assigns association data to the object in response to any commonality between any of the characteristic sensing data and any of the characteristic record data. Assigning the association data to the object includes associating the object with similar or most similar inventory units based on a ranking factor, and determining whether the similarity ranking with the next closest object is high enough to process the first inventory unit via the at least one programmable motion device. If the similarity ranking is too low, the object classification system redirects the object to an exception bin. as well as An assignment system for assigning programmable motion device handling parameters to the first inventory unit of the object based on the associated data, the assignment system including a workflow management system and a separate operable controller, the programmable motion device handling parameters including any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position for engaging the object with the end effector, and placement in which the end effector holds the object during gripping.

10. The object classification system of claim 9, wherein a person is permitted to input either additional characteristic perception data or additional handling parameter input data into the object classification system.

11. The object classification system of claim 9, wherein the object is configured to have multiple similar objects in a box, and wherein the box includes a box identification mark thereon, the box identification mark being sensed by a box sensing system to generate box sensing data, the box sensing data being provided to the object classification system.

12. The object classification system of claim 9, wherein the programmable motion device handling parameters include data regarding whether the object is acceptable for processing by the programmable motion device.

13. The object classification system of claim 9, wherein either previously recorded characteristic perception data or transport parameter input data of different previously classified objects is accessed by the assignment system to help generate the programmable motion device transport parameters of the objects.

14. The object classification system of any one of claims 9 to 13, wherein the programmable motion device handling parameters include data regarding transport instructions for transporting the object to a selected programmable motion device among the plurality of programmable motion devices.

15. The object classification system of any one of claims 9 to 13, wherein the object processing system is used in conjunction with the object classification system, and the object classification system communicates with a plurality of other object processing systems for processing objects.

16. The object classification system of claim 15, wherein the other object processing system is capable of processing objects with specific handling parameters.

17. The object sorting system of claim 16, wherein the dedicated handling parameters include any one of size, weight, placement authority, or arrangement authority.

18. An object classification system for an object processing system including at least one programmable motion device, the object classification system comprising: A tag-sensing system, the tag-sensing system being used to provide a first inventory unit for identifying objects to be processed; An analysis system, the analysis system including at least one feature sensing system for providing feature sensing data about the object to be processed, the feature sensing data including data related to any one of weight, height, width, length, centroid, object description, object category and at least one image; A handling parameter input system is used to obtain handling parameter input data about the object, the handling parameter input data including data related to any of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid. A non-transitory medium for storing the characteristic sensing data and the handling parameter input data associated with the first inventory unit of the object; An association system comprising an object information database having characteristic record data, the characteristic record data including data relating to any one of weight, height, width, length, centroid, object description, object category, and at least one image of each of a plurality of objects, the association system assigning association data to the object in response to commonality between any of the characteristic perception data and any of the characteristic record data, wherein assigning the association data to the object includes associating the object with similar or most similar inventory units based on a ranking factor, and determining whether the similarity ranking with the next closest object is high enough to process the first inventory unit by the at least one programmable motion device; if the similarity ranking is too low, the object classification system redirects the object to an exception bin. as well as An assignment system is used to assign programmable motion device handling parameters to the first stock unit of the object based on the associated data. The programmable motion device handling parameters include any one of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position for engaging the object with the end effector, and placement in which the end effector holds the object during gripping.

19. The object classification system of claim 18, wherein a person is permitted to input either additional characteristic perception data or additional handling parameter input data into the object information database.

20. The object classification system of any one of claims 18 to 19, wherein the objects are configured to have multiple similar objects in a box, and wherein the box includes a box identification mark thereon, the box identification mark being sensed by a box sensing system to generate box sensing data, the box sensing data being provided to the object classification system.

21. The object classification system of claim 18, wherein the programmable motion device handling parameters include data regarding whether the object is acceptable for processing by the programmable motion device.

22. The object classification system of claim 18, wherein either previously recorded characteristic perception data or transport parameter input data of different previously classified objects is accessed by the assignment system to help generate the programmable motion device transport parameters of the objects.

23. The object classification system of claim 18, wherein the programmable motion device handling parameters include data regarding transport instructions for transporting the object to a selected programmable motion device among a plurality of programmable motion devices.

24. The object classification system of claim 18, wherein the object processing system is used in conjunction with the object classification system, and the object classification system communicates with a plurality of other object processing systems for processing objects.

25. The object classification system of claim 24, wherein the other object processing system includes the ability to process objects with specific handling parameters.

26. The object sorting system of claim 25, wherein the dedicated handling parameters include any one of size, weight, placement authority, or arrangement authority.

27. A method for operating an object generalization system for an object processing system including at least one programmable motion device, the method comprising: Provides a first stock unit for identifying objects to be processed; Provide characteristic-aware data about the object to be processed, the characteristic-aware data including data related to any one of weight, height, width, length, centroid, object description, object category and at least one image; Obtain handling parameter input data for the object, the handling parameter input data including data related to any of the following: whether the object can roll after placement, whether the object is fragile, whether the object is stackable, whether the object is crushable, whether the object is deformable, whether the object is too thin to handle, whether the object includes glass, and whether the object is non-rigid; The characteristic sensing data and the handling parameter input data associated with the first inventory unit of the object are stored; Access an object information database with characteristic record data, the characteristic record data including data related to any one of weight, height, width, length, centroid, object description, object category, and at least one image of each of a plurality of objects; In response to the commonality between any of the characteristic-aware data and any of the characteristic-recorded data, association data is assigned to the object, wherein assigning the association data to the object includes associating the object with similar or most similar inventory units based on a ranking factor, and determining whether the similarity ranking with the next closest object is high enough to process the first inventory unit by the at least one programmable motion device; if the similarity ranking is too low, the object induction system redirects the object to the exception bin. as well as Based on the associated data, programmable motion device handling parameters are assigned to the first stock unit of the object. The programmable motion device handling parameters include any of the following: vacuum pressure at the end effector, size of the vacuum suction cup at the end effector, maximum moving speed of the end effector, maximum angular acceleration of the end effector, maximum linear acceleration of the end effector, position where the object engages with the end effector, and placement where the end effector holds the object during gripping.

28. The method of claim 27, wherein the method further comprises allowing personnel to input either additional characteristic sensing data or additional handling parameter input data into the object information database.

29. The method of any one of claims 27 to 28, wherein the object is configured to have a plurality of similar objects in a box, and wherein the box includes a box identification mark thereon, the box identification mark being sensed by a box sensing system to generate box sensing data, the box sensing data being provided to the object induction system.

30. The method of any one of claims 27 to 28, wherein the programmable motion device handling parameters include data regarding whether the object is acceptable for processing by the programmable motion device.

Images