WO2010096850A1

WO2010096850A1 - Method and apparatus for the robot-controlled gripping and moving of objects

Info

Publication number: WO2010096850A1
Application number: PCT/AT2010/000055
Authority: WO
Inventors: Gerhard Kopf; Markus Ikeda; Andreas Pichler
Original assignee: Ih Tech Sondermaschinenbau U. Instandhaltung Gmbh; Profactor Gmbh
Priority date: 2009-02-26
Filing date: 2010-02-26
Publication date: 2010-09-02
Also published as: AT11337U1; DE112010001252A5

Abstract

The invention relates to a method and to an apparatus (1) for the sequential robot-controlled gripping and moving of objects (2) from a plurality of objects (2) disposed in a defined spatial region (3), comprising a device (7) for optically detecting the spatial region (3) with the objects (2) disposed therein, a memory (10) for storing models of the objects (2) and the possible gripping points (9) thereof, a device (8) for identifying the objects (2) and the positions thereof on the basis of the data stored in the memory (10), and determining the collectivity of potential gripping points (9) in the spatial region (3) and the sorting thereof according to relevance so as to determine the object (2) to be preferably gripped and having preferred gripping points (9), and comprising a robot-controlled gripper (5) for gripping the object (2) and moving the object (2) to a defined target location (4). In order to create a method and/or an apparatus (1) that function optimally and are flexible, a simulation apparatus (11) is provided for simulating the gripping and moving process of at least the next defined object (2) prior to carrying the process out and for evaluating the simulation and deciding whether the gripping and moving process of at least the next object (2) is carried out or at least the next object (2) having different gripping points (9) or another preferred object (2) to be gripped having preferred gripping points (9) is determined.

Description

Method and device for robot-controlled gripping and moving of objects

The invention relates to a method for the sequential robot - controlled gripping and moving of objects from a plurality of objects arranged in a defined spatial area, the spatial area with the objects arranged therein being optically detected, the objects being identified on the basis of stored models of the objects and their position in the object is determined on the basis of the potential gripping points stored for all objects, the gripping points of the objects are sorted according to their relevance and due to the relevance of the gripping points a preferably to be gripped object is determined with preferred gripping points, this Object is grabbed by a robot-controlled gripper and moved to a defined destination, and then determines the next preferred object to be gripped with preferred gripping points and then this object grasped and moved to the defined destination we d.

The invention further relates to an apparatus for carrying out a method for the sequential robot-controlled gripping and moving of objects from a plurality of objects arranged in a defined spatial area, with a device for optically detecting the spatial area with the objects arranged therein, a memory for Depositing models of the objects and their possible gripping points, a device for identifying the objects and their location on the basis of the data stored in memory, and determining the totality of possible gripping points in the spatial area and their sorting according to their relevance for determining the preferred to be gripped Object with preferred grip points, and with a robotic gripper for gripping the object and moving the object to a defined destination.

The integration of robot systems into production lines involves a great deal of time and logistics. Industrial robots execute pre-programmed movements. Therefore, a great deal of effort is required in the presentation of the objects to be moved, since the objects to be manipulated are singled out and precisely defined. well-defined destinations must be prepared in well-defined locations before they become tangible for robotics and automation. These are currently major obstacles on the way to so-called adaptive production, which reduce the economics of customer customization and rapid innovation. Therefore, robots are currently used for only about 25% of all purposeful applications.

A variety of documents are concerned with methods and apparatus for sequentially robotic gripping and moving objects.

Typically, a subject robotic system consists of a spatial domain optically sensing device in which the objects to be grasped and moving are located, at least one database for storing important process data, means for computing and controlling the robot, and finally the robot for gripping and moving the obj ekte.

For example, WO 2007/046763 A1 describes a method and a device for locating and gripping objects arranged on a base, wherein improvements are to be achieved by a special three-dimensional data acquisition by means of a light-section method.

EP 951 968 A2 describes a device of the type in question, wherein the objects to be gripped are separated by means of a vibrating mechanism, so that the isolated objects can be better gripped. The gripping process itself and the planning of a gripping hierarchy are not dealt with in this object.

EP 1 428 634 A2 describes a mechanism for removing objects from a container using a three-dimensional optical detection system.

EP 1 589 483 A2 describes a gripping system in which the objects to be gripped are optically detected and due to stored parameters about the objects to be gripped a certain se GreifStrategie is fed to the robot.

EP 1 385 122 B1 shows an apparatus for taking objects with an image recording device for detecting the objects, memory devices for storing information about the objects and devices for determining the priority of the extraction of the objects and for forwarding to the robot for removing the objects.

The known methods and devices are usually very complex and inflexible, so that they are used only in processes with a very high number of objects and only for certain objects. An application of known methods and devices for use in smaller amounts of objects usually fails because of the expense of programming the procedures and procedures. A rapid and rapid adaptation to new objects to be gripped and moved is usually possible only with great effort and expense using conventional methods and devices.

The object of the present invention is therefore to provide a method and apparatus mentioned above, which allow optimal recognition of the objects and optimal determination of a gripping hierarchy, so that the objects can be moved quickly from the spatial area to corresponding target areas. Furthermore, the objective method and the subject device should be characterized by a high degree of flexibility and easy adaptability to new objects. Disadvantages of known methods and devices should be avoided or at least reduced.

The object according to the invention is achieved by an above-mentioned method for the sequential robot-controlled gripping and moving of objects, in which the gripping and moving process of at least the next specific object is simulated before the execution, the simulation is evaluated and a decision is made based on the evaluation of the simulation, whether the gripping and moving process of at least the next object is performed or at least the next object with other gripping points or another object to be gripped preferably with gripping is determined. The inventive method is characterized in that the planning of the gripping process based on the object recognition taking into account stored data on the objects to be detected, said before the execution of the gripping and moving process simulated this process for at least the next planned object, the simulation evaluated and It is only decided after that whether the gripping and moving operation is performed on this object at the certain gripping points, or the gripping and moving process is simulated at other gripping points of the same object or another object with preferred gripping points and then performed according to the evaluation. By simulating the gripping and moving processes even nonsensical or hazardous to the robot-controlled gripper movements can be prevented, and thus the process can be performed quickly and safely. By changing the stored models of the objects as well as the possible gripping points of these objects, the process can be quickly and inexpensively adapted to new parts. As a result, use of the method can also be carried out economically in smaller plants and for a smaller number of objects to be moved.

According to a further feature of the method according to the invention, in addition to the models of all objects and their possible gripping points, additional parameters of the objects are stored and the next object to be gripped and its preferred grip points are also determined on the basis of these additional parameters. Through such additional parameters stored in a corresponding database, the grasp hierarchy can be determined more quickly and, above all, more efficiently. Such additional parameters of the objects include, for example, position properties of the objects, information about the material of the objects or much more.

Advantageously, when sorting the possible gripping points according to their relevance, the height of the gripping point in the spatial area is taken into account and the gripping points with the greatest height are preferred. This ensures that the overhead parts in the spatial area are removed first and thus the arrangement of the objects by the removal of an object not changed much. Such a change in the three-dimensional arrangement of the objects would again have to be influenced separately and the grip planning and simulation of the same would be renewed, which could lead to delays.

Furthermore, it is advantageous if the gripping points of individual objects are preferred when sorting the possible gripping points according to their relevance. This ensures that individual objects are grasped and moved first.

When evaluating the simulation of the gripping and moving process of an object, preferably stored data of the geometry and kinematics of the robot-controlled gripper are taken into account. With this stored data, the gripping strategy can be planned taking into account the optimum geometry and kinematics of the robot-controlled gripper or the entire robot system. When changing the robot system, for example, when changing the gripper, the corresponding data in the database can be changed in a simple manner and thus made a quick adjustment.

When evaluating the simulation of the gripping and moving process of an object, in addition preferably deposited data of the geometry of the environment of the spatial area are taken into account. This prevents collisions between the object or robot system and parts of the environment and thus delays in the gripping and moving process can be avoided.

According to a further feature of the invention, it is provided that unidentifiable objects or unidentifiable layers of objects are detected and corrected. This can also increase the efficiency of gripping planning.

If the optical detection of the spatial area is changed with the objects arranged therein, for example, the exposure is changed, the Griffplanung can be supported. Certain changes in the optical detection can be used to optimize the recognition of the objects and, as a result, the gripping planning. Depending on spatial conditions and on the application, certain preferred regions can be defined within the spatial region with the objects arranged therein, and gripping points of objects lying within these preferential regions can be preferred. By defining such preferential regions, for example, certain situations in a manufacturing process or the like. Be taken into account and finally an increase in efficiency in the gripping planning can be achieved.

Preferably, different paths for the movements of an object to the destination are determined and sorted according to predetermined criteria. As a result, the flexibility in gripping planning can be increased.

Preferably, the determined and performed gripping and moving operations of the objects are stored and possibly used in subsequent procedures. In this way, a kind of learning system can be created using historical data.

According to a further feature of the invention it is provided that the speed of the gripping and moving operations of the objects is changed. By adjusting the speed when gripping and moving the objects, certain situations can be influenced, thereby achieving an optimum adaptation of the speed of the gripping and moving operations.

The object according to the invention is also achieved by an abovementioned device, wherein a simulation device for simulating the gripping and moving process of at least the next specific object before carrying it out and for evaluating the simulation and deciding whether the gripping and moving process at least the next object is performed or at least the next object with other gripping points or another object to be gripped preferably determined with preferred gripping points is provided. For the advantages of such an embodiment, reference is made to the above description of the method. The simulation device is preferably formed by a computer. The optical detection device can be formed by a two-dimensional image processing system, in particular a camera, or else by a three-dimensional image processing system.

Preferably, a memory is provided for storing additional parameters of the objects in addition to the models of the objects and their possible gripping points. These additional parameters, like the models of the objects and the possible gripping points of the objects, can be quickly and easily stored or changed in the memory.

Advantageously, a memory for storing data of the geometry and kinematics of the robot-controlled gripper is also provided. Corresponding data about the geometry and the kinematics of the robot-controlled gripper or entire robot system are stored in this memory.

Likewise, a memory for storing data of the geometry of the environment of the spatial area may be provided, so that these data can be taken into account in the planning of the gripping strategy.

To adapt and optimize the speed of the movements, a device for changing the speed of the gripping and moving operations of the robot-controlled gripper can be provided.

The present invention will be explained in more detail with reference to the accompanying drawings. Show:

Fig. 1 is a schematic flow diagram of the method according to the invention;

Fig. 2 is a schematic block diagram of the device according to the invention;

3a and 3b show two views of an object with arranged and visualized gripping points; FIGS. 4a and 4b show two views of the object according to FIGS. 3a and 3b with grippers arranged at all possible gripping points; and

Fig. 5 shows the situation of FIG. 4, in which the gripping positions are shown only at those gripping points of the object, which are reachable without collision.

1 shows a flow chart of the method according to the invention for the sequential robot-controlled gripping and moving of objects from a multiplicity of objects arranged in a defined spatial area. After starting the method according to block 100, an optical detection of the spatial area with the objects arranged therein takes place in accordance with block 101. This optical detection can be done with corresponding two-dimensional or three-dimensional image processing systems. For the object recognition and position recognition of the objects, data is accessed in memories or databases 102 to 104 which contain stored two-dimensional or three-dimensional models of the objects as well as their possible gripping points. In addition, additional parameters of the objects can be stored in corresponding memories or databases. In addition, data of the geometry and kinematics of the robot-controlled gripper or robot system can be stored and taken into account in the object recognition. Finally, data of the geometry and environment of the spatial area can be stored and taken into account in the object recognition. After recognition of the objects and their location, the spatial representation of the plurality of objects in the spatial area takes place in accordance with block 105. In block 106, all possible grip points are determined. Finally, according to block 107, the determination of the relevance of the gripping points of the objects takes place and, finally, according to block 108, the determination of an object preferably to be gripped with preferred gripping points, whereby appropriate strategy parameters (block 109) can be used to determine the gripping strategy. After determining the object preferably to be gripped with preferred gripping points, according to block 110, the simulation of the gripping and moving process of at least the next specific object and an evaluation of this simulation are carried out. The decision according to block 111 determines whether the simulation is found to be in order, whereupon, according to block 112, the gripping and moving operation is performed and returned to the object recognition according to block 101 again. If, for example, the simulation is rejected due to a possible collision of the robot or object during the movement process, the system returns to block 106 and determines the object with new gripping points or another object with preferred gripping points as the next object to be gripped and moved.

FIG. 2 shows a block diagram of a possible embodiment of a device 1 for carrying out a method for the sequential robot-controlled gripping and moving of objects

2 of a variety of in a defined spatial area

3 arranged objects 2. The device 1 is used to move from a plurality of objects 2 individual objects 2 from the spatial area 3 to a defined destination 4. For this purpose, a robot-controlled gripper 5 is used. The term "gripper" covers all possible handling devices, such as the jaw grippers illustrated, but also suction devices, devices using magnets or much more.As a possible destination 4, for example, a conveyor belt can be used which integrates the separated objects 2 into a process sequence The robot-controlled gripper 5 is connected to a control device 6, which is formed for example by a computer 7. A device 7 for the optical detection of the spatial area 3 with the objects 2 arranged therein is connected to a device 8 for identifying the objects 2 and their position such that the objects 2 can be recognized and displayed within the spatial area 3. For object recognition, models of the objects 2 and their possible gripping points 9, which are stored in a memory 10, are accessed respective Objects 2 are more quickly identified in the spatial area 3 and their location can be determined. In addition to the two-dimensional or three-dimensional models of the objects 2 and the possible gripping points 9, additional parameters can also be stored in the same memory 10 or a further memory. The device 8 for identifying the objects 2 and displaying their location also serves for determining the totality of possible gripping points 9 in the spatial area 3 and their sorting according to their relevance for determining the object 2 preferably to be gripped with preferred gripping points 9. According to the invention, a simulation device 11 is provided for simulating the gripping and moving process. In this simulation device 11, the gripping and moving process of at least the next specific object 2 is simulated prior to its execution and an evaluation of the simulation made and finally decided whether the gripping and moving operation is performed by the robot-controlled gripper 5 or the same particular object 2 with other gripping points 9 or another object 2 with preferred gripping points 9 from the spatial area 3 is selected. Of course, the control device 6, the device 8 for identifying the objects 2 and the simulation device 11 can be accommodated in a common computer unit 12 as shown.

For the device 7 for the optical detection of the spatial area 3 and the objects 2 arranged therein, a two-dimensional or three-dimensional image processing system can be used.

In a further memory 13, data of the geometry and kinematics of the robot-controlled gripper 5 can be stored and used in the selection of the gripping strategy by the device 8 for identifying the objects 2 and the simulation device 11. In a further memory 14 and data of the geometry and environment of the spatial area 3 may be deposited in order to exclude any collisions during the gripping and moving process can. Of course, the various memories 10, 13, 14 may be formed by a single memory or a database. The present device 1 is characterized by particularly high flexibility, since the device 1 can be adapted to changed conditions and changed objects 2 quickly and easily.

Finally, FIGS. 3 a and 3 b show two views of a three-dimensional object 2 with grippers 5 arranged at all possible gripping points 9. 4a and 4b show the object 2 shown in FIGS. 3a and 3b with grippers 5 arranged thereon at all possible gripping points 9.

In Fig. 5, only the gripping points 9 are listed on the objects 2, which would not lead to collisions.

Claims

Claims :

Method for the sequential robot-controlled gripping and moving of objects (2) from a multiplicity of objects (2) arranged in a defined spatial area (3), wherein the spatial area (3) with the objects (2) arranged therein is optically detected on the basis of stored models of the objects (2) the objects (2) identified and their location in the spatial area (3) is determined, based on the objects (2) stored possible gripping points (9) the totality of possible gripping points (9) of all Objects (2) is determined, the gripping points (9) of the objects (2) are sorted according to their relevance and due to the relevance of the gripping points (9) a preferably to be gripped object (2) with preferred gripping points (9) is determined, this object (2) is gripped by a robot-controlled gripper (5) and moved to a defined destination (4), and thereafter the next object to be gripped (2) with preferred gripping points (9) determined and dana This object (2) is gripped and moved to the defined destination (4), characterized in that the gripping and moving process of at least the next specific object (2) is simulated before the execution, the simulation is evaluated and based on the evaluation of the simulation It is decided whether the gripping and moving process of at least the next object (2) is performed or at least the next object (2) with other gripping points (9) or another object (2) to be gripped preferably with preferred gripping points (9) is determined.

2. The method according to claim 1, characterized in that in addition to the models of all objects (2) and their possible gripping points (9) additional parameters of the objects (2) are deposited and the next preferably to be gripped object (2) and its preferred grip points ( 9) can also be determined on the basis of these additional parameters.

3. The method according to claim 1 or 2, characterized in that when sorting the possible gripping points (9) according to their relevance, the height of the gripping point (9) in the spatial area (3) is taken into account and the gripping points (9) with the greatest height preferred become.

4. The method according to any one of claims 1 to 3, characterized in that when sorting the possible gripping points (9) according to their relevance, the gripping points (9) of individual objects (2) are preferred.

5. The method according to any one of claims 1 to 4, characterized in that in the evaluation of the simulation of the gripping and moving operation of an object (2) stored data of the geometry and kinematics of the robot-controlled gripper (5) are taken into account.

6. The method according to any one of claims 1 to 5, characterized in that in the evaluation of the simulation of the gripping and moving operation of an object (2) stored data of the geometry of the environment of the spatial area (3) are taken into account.

7. The method according to any one of claims 1 to 6, characterized in that unidentifiable objects (2) or unidentifi- zierbare layers of objects (2) are detected and corrected.

8. The method according to any one of claims 1 to 7, characterized in that the optical detection of the spatial area (3) with the objects arranged therein (2) is changed, for example, the exposure is changed.

9. The method according to any one of claims 1 to 8, characterized in that within the spatial area (3) certain preferential regions are defined and lying within these preferential regions gripping points (9) of objects (2) are preferred.

10. The method according to any one of claims 1 to 9, characterized in that different paths for the movements of an object (2) to the destination point (4) are determined and sorted according to predetermined criteria.

11. The method according to any one of claims 1 to 10, characterized in that the specific and carried out gripping and Be wegungsvorgänge the objects (2) stored and possibly used in subsequent procedures.

12. The method according to any one of claims 1 to 11, characterized in that the speed of the gripping and moving operations of the objects (2) is changed.

13. Apparatus (1) for carrying out a method for the sequential robot-controlled gripping and moving of objects (2) from a multiplicity of objects (2) arranged in a defined spatial area (3), comprising means (7) for optical detection of the spatial Area (3) with the objects (2) arranged therein, a memory (10) for depositing models of the objects (2) and their possible gripping points (9), means (8) for identifying the objects (2) and their position based on the data stored in the memory (10), and the determination of the totality of possible gripping points (9) in the spatial area (3) and their sorting according to their relevance for determining the preferably to be gripped object (2) with preferred gripping points (9), and with a robot-controlled gripper (5) for gripping the object (2) and moving the object (2) to a defined destination (4), characterized in that a simulation device (11) for simulation the gripping and moving operation of at least the next specific object (2) prior to its execution and for evaluating the simulation and deciding whether the gripping and moving operation performed at least the next object (2) or at least the next object (2) with other gripping points (9) or another preferably to be gripped object (2) is determined with preferred gripping points (9) is provided.

14. Device (1) according to claim 13, characterized in that the optical detection device (7) is formed by a 2D image processing system, in particular a camera.

15. Device (1) according to claim 13, characterized in that the optical detection device (7) is formed by an SD image processing system.

16. Device (1) according to any one of claims 13 to 15, characterized in that a memory is provided for storing additional parameters of the objects (2) in addition to the models of the objects (2) and their possible gripping points (9).

17. Device (1) according to any one of claims 13 to 16, characterized in that a memory (13) for storing data of the geometry and kinematics of the robot-controlled gripper (5) is provided.

18. Device (1) according to any one of claims 13 to 17, characterized in that a memory (14) for storing data of the geometry of the environment of the spatial area (3) is provided.

19. Device (1) according to any one of claims 13 to 18, characterized in that a device for changing the speed of the gripping and moving operations of the robot-controlled gripper (5) is provided.