Object grasping method and program

Abstract

To provide a method for gripping an object that is further robust against measurement errors and can grip an object stably, and to provide a program.SOLUTION: A method for gripping an object includes the steps of: determining a center of a first object assumed to be gripped by an end effector relating to a plurality of gripping postures that the end effector can take; measuring a shape of a second object on the basis of an image capturing the second object from a certain one view point; selecting a gripping posture when the end effector grips the second object from the plurality of gripping postures on the basis of the centers of the second object and the first object; and contracting part of the shape of the second object when the end effector grips the second object while taking the selected gripping posture.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an object grasping method and a program. [Background technology]

[0002] Techniques for making a robot grasp an object have been studied (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6476358 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional technologies measure the shape of an object and then grasp it using a gripper with low shape dependency, search for a possible graspable posture and grasp it, or decide on a grasp method and grasp it by continuing to move the joints at a predetermined speed rate. However, conventional technologies have problems such as being able to grasp an object but not being able to manipulate it, taking a very long time to search for a possible graspable posture, low robustness against shape errors, and being able to grip an object but having difficulty grasping it with the fingertips.

[0005] The present invention has been made in consideration of these circumstances, and one of its objectives is to provide an object grasping method and program that are more robust against measurement errors and can grasp an object stably. [Means for solving the problem]

[0006] The object grasping method and program according to the present invention employ the following configuration. (1) A first aspect of the present invention is an object grasping method including the steps of: determining a center of a first object that is expected to be grasped by an end effector for each of a plurality of grasping postures that the end effector can take; measuring a shape of the second object based on an image of the second object captured from at least one viewpoint; selecting one grasping posture from the plurality of grasping postures based on the center of the second object whose shape has been measured and the center of the first object when the end effector is to grasp the second object; and, when the end effector is to grasp the second object in the selected grasping posture, reducing a portion of the measured shape of the second object that is in a blind spot when viewed from the viewpoint and through which force is applied from the end effector to the second object.

[0007] (2) A second aspect of the present invention is a method for selecting a gripping posture in the first aspect, wherein the step of selecting a gripping posture includes estimating the center of the second object whose shape has been measured, and selecting, from the plurality of gripping postures, a gripping posture associated with the first object having the same center as the estimated center of the second object.

[0008] (3) A third aspect of the present invention is the first or second aspect, wherein the plurality of gripping positions includes a first gripping position in which the side surface of a horizontal cylinder is gripped.

[0009] (4) A fourth aspect of the present invention is the third aspect, wherein the first grasping posture includes a constraint condition that directs the force applied from the end effector to the side surface of the horizontal cylinder when the end effector contacts the side surface toward the central axis, which is the longitudinal direction of the horizontal cylinder.

[0010] (5) A fifth aspect of the present invention is any one of the first to fourth aspects, wherein the plurality of gripping postures includes a second gripping posture for gripping a sphere.

[0011] (6) A sixth aspect of the present invention is the fifth aspect, wherein the second grasping posture includes a constraint condition that directs the force applied from the end effector to the surface of the sphere when the end effector contacts the surface of the sphere toward the center point of the sphere.

[0012] (7) A seventh aspect of the present invention is any one of the first to sixth aspects, wherein the plurality of gripping positions includes a third gripping position in which the end of a vertical cylinder is gripped.

[0013] (8) An eighth aspect of the present invention is the seventh aspect, wherein the third grasping posture includes a constraint condition that directs the force applied from the end effector to the end of the vertical cylinder when the end effector contacts the end toward the central axis, which is the longitudinal direction of the vertical cylinder.

[0014] (9) A ninth aspect of the present invention is any one of the first to eighth aspects, wherein the plurality of gripping positions includes a fourth gripping position in which an end of a rectangular parallelepiped is gripped.

[0015] (10) A tenth aspect of the present invention is the ninth aspect, wherein the fourth grasping posture includes a constraint condition that causes the force applied from the end effector to the end of the rectangular parallelepiped when the end effector contacts the end to be parallel to the central axis, which is the longitudinal direction of the rectangular parallelepiped.

[0016] (11) An eleventh aspect of the present invention is a program for causing a computer to execute the following steps: determining the center of a first object that is expected to be grasped by the end effector for each of a plurality of grasping postures that the end effector can take; measuring the shape of the second object based on an image of the second object captured from at least one viewpoint; selecting one grasping posture from the plurality of grasping postures when the end effector is to grasp the second object based on the center of the second object whose shape has been measured and the center of the first object; and reducing a part of the measured shape of the second object that is in a blind spot when viewed from the viewpoint and where force is applied from the end effector to the second object when the end effector is to grasp the second object in the selected grasping posture. [Effects of the Invention]

[0017] According to the above aspect, it is possible to grip an object more robustly against measurement errors and more stably. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a diagram schematically illustrating the appearance of a robot 10 according to an embodiment. [Figure 2] 1 is a configuration diagram of a robot 10 and a control device 100 according to an embodiment. [Figure 3] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 4] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 5] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 6] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 7] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 8] FIG. 1 is a diagram illustrating an example of a grip taxonomy. [Figure 9]10 is a flowchart showing the flow of a series of processes by a processing unit 120 according to the embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of a method for selecting a gripping posture. [Figure 11] FIG. 10 is a diagram illustrating an example of a method for selecting a gripping posture. [Figure 12] FIG. 10 is a diagram illustrating an example of a method for selecting a gripping posture. [Figure 13] 10A and 10B are diagrams for explaining another example of a method for selecting a gripping posture. [Figure 14] 10A and 10B are diagrams for explaining another example of a method for selecting a gripping posture. [Figure 15] 10A and 10B are diagrams for explaining another example of a method for selecting a gripping posture. [Figure 16] 10A and 10B are diagrams for explaining a method for correcting the shape of an object OB. [Figure 17] 10A and 10B are diagrams for explaining a method for correcting the shape of an object OB. [Figure 18] 10A and 10B are diagrams for explaining a method for correcting the shape of an object OB. DETAILED DESCRIPTION OF THE INVENTION

[0019] Hereinafter, an embodiment of an object gripping method and a program according to the present invention will be described with reference to the drawings.

[0020] [Robot appearance] FIG. 1 is a diagram schematically illustrating the appearance of a robot 10 according to an embodiment. The robot 10 is typically a humanoid robot capable of gripping an object OB using an end effector, but is not limited to this and may be any type of robot capable of gripping an object OB. The end effector is also referred to as a robot hand. Therefore, in the following description, the end effector may be simply referred to as a "hand." The hand of the robot 10 may have, for example, four fingers provided as grippers.

[0021] The robot 10 is equipped with, for example, a camera 11 for imaging the external environment as seen by the robot 10, and a control device 100 for controlling the operation of the robot 10, and performs the desired task according to the actions determined by the control device 100.

[0022] The task may be, for example, grabbing the object OB with one hand, transferring the object OB to the other hand, or moving the object OB. However, the task is not limited to these, and any task may be set.

[0023] The camera 11 is installed in a part of the body (typically the head) of the robot 10. The camera 11 may be, for example, a depth camera (3D camera). For example, the camera 11 captures an image of the object OB and transmits image data of the object OB to the control device 100.

[0024] The control device 100 performs autonomous control so that the robot 10 behaves autonomously, for example. The control device 100 is typically mounted on the robot 10 and directly controls the robot 10. Alternatively, the control device 100 may be installed in a remote location far away from the robot 10 and control the robot 10 remotely via a network NW. The network NW includes a LAN (Local Area Network) and a WAN (Wide Area Network), etc.

[0025] [Robot and control device configuration] 2 is a configuration diagram of a robot 10 and a control device 100 according to the embodiment. In addition to the camera 11 described above, the robot 10 further includes an actuator 12, a state sensor 13, and an operation control unit .

[0026] The actuators 12 are controlled by the motion control unit 14 to drive the various parts (arms, fingers, legs, head, torso, waist, etc.) of the robot 10. The actuators 12 include, for example, electromagnetic motors, gears, artificial muscles, etc.

[0027] The state sensor 13 is a sensor that detects the state (for example, joint angle, angular velocity, torque, etc.) of the robot 10. The state sensor 13 includes, for example, a rotary encoder that detects the degree of rotation of the joint of the robot 10, a tension sensor that detects the tension of the wire that rotates the joint, a torque sensor that detects the torque applied to the joint axis, an acceleration sensor and a gyro sensor that detect the posture of the robot 10, etc.

[0028] The operation control unit 14 controls the actuator 12 based on the control command generated by the control device 100 .

[0029] The control device 100 includes, for example, a communication interface 110, a processing unit 120, and a storage unit .

[0030] The communication interface 110 communicates with an external device via a network NW, and with the robot 10 via a communication line such as a bus. The external device is, for example, a terminal device (such as a smartphone or a personal computer) that can be used by a user to request the robot 10 to execute a task. The communication interface 110 includes, for example, a wireless communication module including a receiver and a transmitter, a NIC (Network Interface Card), and the like.

[0031] The processing unit 120 includes, for example, an acquisition unit 121, a shape measurement unit 122, a grip posture determination unit 123, a shape correction unit 124, a command generation unit 125, and a communication control unit 126.

[0032] The components of the processing unit 120 are realized by, for example, a central processing unit (CPU) or a graphics processing unit (GPU) executing programs or instructions stored in the storage unit 130. Some or all of these components may be realized by hardware such as a large-scale integration (LSI), an application-specific integrated circuit (ASIC), or a field-programmable gate array (FPGA), or may be realized by a combination of software and hardware.

[0033] The storage unit 130 is realized by, for example, a hard disk drive (HDD), a flash memory, an electrically erasable programmable read-only memory (EEPROM), a read-only memory (ROM), a random access memory (RAM), etc. The storage unit 130 stores a grip taxonomy in addition to firmware, application programs, etc.

[0034] The grasping taxonomy is a database in which multiple grasping postures that an end effector can take are classified and organized, and the relationship between each classified grasping posture and the center of a reference object OBref that is expected to be grasped by the end effector is predefined. The "center" refers to a central axis or central point, which will be described later. The grasping taxonomy may be installed in the storage unit 130 from an external device via the network NW, or may be installed in the storage unit 130 from a portable storage medium connected to a drive device of the control device 100. The reference object OBref in the grasping taxonomy is an example of a "first object."

[0035] Figures 3 to 8 show an example of a grasping taxonomy. Figure 3 shows a grasping posture called "medium-wrap," which systematizes grasping a "horizontal cylinder" as the reference object OBref. For example, if the length direction of a horizontal cylinder is the Y axis (if the horizontal cylinder extends in the Y-axis direction), the central axis of the horizontal cylinder is the Y axis. Medium-wrap defines grasping the side of a horizontal cylinder using all fingers (grippers) along the X-axis or Z-axis, which are perpendicular to the Y axis, the central axis of the horizontal cylinder. In other words, when the fingertips contact the side of the horizontal cylinder, the force applied from the fingertips to the side (hereinafter referred to as contact force) is directed toward the Y axis, the central axis of the horizontal cylinder. Medium-wrap is an example of a "first grasping posture."

[0036] Figure 4 shows a grasping posture called "power-sphere," which organizes the grasping of a "sphere" as the reference object OBref. Power-sphere defines grasping a sphere so that the contact forces of all fingers are directed toward the center of the sphere. In other words, when the reference object OBref is a sphere, there are an infinite number of centerlines. Therefore, power-sphere selects the centerline that allows grasping with the least change from the current grasping posture of the hand, and grasps the sphere so that the contact forces are directed toward that centerline. Power-sphere is an example of a "second grasping posture."

[0037] Figure 5 shows a grasping posture called "tripod," which organizes the grasping of a "vertical cylinder" as a reference object OBref. For example, if the length direction of the vertical cylinder is the Z axis (if the vertical cylinder extends in the Z-axis direction), the central axis of the vertical cylinder is the Z axis. In this case, tripod defines grasping or gripping the end of the vertical cylinder with three fingers so that the contact forces of each of the three fingers are directed along the Z axis. Tripod is an example of a "third grasping posture."

[0038] Figure 6 shows a grasping posture called the "palmar pinch," which is organized as grasping a rectangular "cuboid" with all sides as the reference object OBref. For example, if the longitudinal direction of the cuboid is the X-axis (if the cuboid extends in the X-axis direction), the central axis of the cuboid is the X-axis. In this case, the palmar pinch is defined as pinching the edge of the cuboid with two fingers so that the contact forces of each of the two fingers are parallel to the X-axis. The palmar pinch is an example of the "fourth grasping posture."

[0039] Figure 7 shows a grasping posture called "prismatic-2fingers." In prismatic-2fingers, similar to palmer-pinch, a rectangular prism is grasped as the reference object OBref. However, prismatic-2fingers targets a wider prism than palmer-pinch. When the central axis of the prism is the X-axis, prismatic-2fingers defines the prism as being pinched by three fingers at the edge of the prism so that the contact forces of each of the three fingers are parallel to the X-axis. Prismatic-2fingers is another example of the "fourth grasping posture."

[0040] Figure 8 shows a grasping posture called "prismatic-3fingers." Prismatic-3fingers, like the palmer pinch and prismatic-2fingers, is structured to grasp a rectangular prism as the reference object OBref. However, prismatic-3fingers targets a wider prism than prismatic-2fingers. When the central axis of the prism is the X-axis, prismatic-3fingers defines the prism as being pinched by the four fingers at the edges so that the contact forces of each of the four fingers are parallel to the X-axis. Prismatic-3fingers is another example of the "fourth grasping posture."

[0041] The grip taxonomy is not limited to the above examples, and gripping postures may also be systematized for reference objects OBref of the same shape, such as triangular prisms and cones.

[0042] [Control device processing flow] The following describes a specific processing flow of the processing unit 120 using a flowchart. Fig. 9 is a flowchart showing a series of processing flows of the processing unit 120 according to the embodiment.

[0043] First, the acquisition unit 121 acquires image data and status data from the robot 10 via the communication interface 110 (step S100).

[0044] The image data is image data generated when the camera 11 captures an image of the object OB. For example, if the camera 11 is installed on the head of the robot 10, the acquisition unit 121 acquires image data when the object OB is captured from the viewpoint of the head of the robot 10.

[0045] The state data includes, for example, the detection values ​​of the state sensors 13 of the robot 10. Specifically, the state data includes the joint angles detected by the rotary encoders, the tension of the joint wires detected by the tension sensors, the torque of the joint shafts detected by the torque sensors, the acceleration of the robot 10 detected by the acceleration sensors, the angular velocity of the robot 10 detected by the gyro sensors, etc.

[0046] Next, the shape measuring unit 122 measures the shape of the object OB based on the image data acquired by the acquiring unit 121 (step S102). "Measuring the shape" may be read as "estimating the shape."

[0047] For example, the shape measurement unit 122 extracts a group of points that indicate the contour of the object OB on the image, and calculates the smallest primitive P that encloses the group of points. A Set the primitive P on the image. Ais a basic shape used in modeling for graphic analysis, and is typically a simple shape such as a rectangular parallelepiped, cylinder, or sphere. A When set, the primitive P A The shape of the object OB is the shape of the object OB.

[0048] The shape measuring unit 122 may treat the point group indicating the contour of the object OB as the shape of the object OB.

[0049] Next, the shape measurement unit 122 estimates the central axis of the object OB whose shape has been measured (step S104).

[0050] For example, the shape measurement unit 122 measures a primitive P A If you set , the primitive P A The longitudinal direction of the object OB is the central axis of the object OB.

[0051] For example, if the shape measurement unit 122 considers the point cloud that indicates the contour of the object OB to be the shape of the object OB, it may search for the longest straight line within the area of ​​the point cloud and use the direction parallel to that longest straight line as the central axis of the object OB.

[0052] Next, the gripping posture determination unit 123 determines the gripping posture of the object OB based on the center of the object OB and the center of the reference object OBref, whose relationship to the gripping posture is defined in the gripping taxonomy (step S106).

[0053] For example, the gripping posture determination unit 123 selects one gripping posture from among a plurality of gripping postures whose relationships are defined with respect to the center of the reference object OBref in the gripping taxonomy, using the center of the object OB as a constraint condition.

[0054] 10 to 12 are diagrams for explaining an example of a method for selecting a gripping posture, in which the shape of the object OB is known to some extent.

[0055] For example, it is assumed that the user knows the shape of the object OB to some extent and provides information about the shape, such as which part of the object OB should be grasped, to the control device 100 using a terminal device. In this case, the shape measurement unit 122 maps a primitive P to the part that the robot 10 should grasp, which is specified by the user. A In the example in Figure 10, P A1 , P A2 , P A3 A primitive can be set in a total of three places. For example, the "cuboid" primitive P A1 In this case, the shape measurement unit 122 calculates the primitive P A1 The longitudinal direction of the object OB is estimated as the central axis of the object OB.

[0056] In response to this, for example, the gripping posture determination unit 123 refers to the gripping taxonomy, selects a reference object OBref having the same central axis as the central axis of the object OB, and selects the gripping posture having a relationship defined to the central axis of this reference object OBref as the gripping posture of the object OB.

[0057] For example, as shown in Figure 11, the "cuboid" primitive P A1 A long and narrow grasp center primitive P has a width that allows two fingers to be placed on one end and one finger to be placed on the other end in a direction perpendicular to the longitudinal direction (the central axis of the object OB). B can be set, the gripping posture determination unit 123 selects the gripping posture "Prismatic 2 Fingers" as the gripping posture for the object OB.

[0058] Grasp center primitive P B is a primitive that should be the center of the grasp that determines where the fingers of the robot 10 will contact the object OB, and is a primitive that is compared with the center of the reference object OBref whose relationship with the grasp posture is defined in the grasp taxonomy described above.

[0059] For example, as shown in Figure 12, the "cuboid" primitive P A1The grip center primitive P of a horizontal cylinder whose length direction is the longitudinal direction (the central axis of the object OB) B can be set, the gripping posture determination unit 123 selects the gripping posture of "medium lap" as the gripping posture of the object OB.

[0060] 13 to 15 are diagrams for explaining another example of a method for selecting a gripping posture, in which the shape of the object OB is unknown.

[0061] For example, when the shape of the object OB is unknown, the shape measurement unit 122 may use polygon fitting or the like to fit a polyhedral primitive P A Then, the shape measurement unit 122 sets the primitive P A The longest line within the region is searched for, and the direction parallel to that longest line is estimated as the central axis of the object OB.

[0062] In response to this, the grip posture determination unit 123 determines the grip center primitive P B If it is possible to set, the grasping posture "Prismatic 2 Fingers" is selected as the grasping posture for the object OB.

[0063] 15, the gripping posture determination unit 123 determines a gripping center primitive P B If it is possible to set the gripping posture "medium wrap", the gripping posture "medium wrap" is selected as the gripping posture for the object OB.

[0064] Returning to the description of the flowchart in Fig. 9, the shape correction unit 124 then corrects the shape of the object OB measured by the shape measurement unit 122 based on the blind spot of the camera 11 and the gripping direction (step S108).

[0065] 16 to 18 are diagrams illustrating a method for correcting the shape of the object OB. FIG. 16 illustrates a state in which the object OB is being grasped in a grasping posture known as "prismatic two fingers." Since the Y axis in the figure is the central axis of the object OB, the object OB is grasped using three fingers so that the contact forces of each of the three fingers are parallel to the X axis, which is perpendicular to the Y axis. Here, if the camera 11 is installed on the head, which is higher than the hand, the back side of the object OB, for example, is in the blind spot of the camera 11, which makes it more likely that errors will occur in the measurement of the object OB. Therefore, the shape correction unit 124 corrects the shape of the object OB, assuming that a measurement error has occurred.

[0066] For example, Figure 17 shows a polyhedral primitive P placed over the object OB. A The figure shows a state in which an object OB whose shape has been measured with higher precision using a laser beam or the like (hereinafter referred to as an object OB whose shape is close to the actual shape) is superimposed on the object OB. As shown in the figure, in the blind spot area of ​​the camera 11 (part R in the figure), the polyhedral primitive P A The shape of the object, i.e., the shape of the object OB, is measured to a large extent.

[0067] In such a case, the shape correction unit 124 reduces a part of the shape of the object OB that is measured as being largely protruding, as shown in Fig. 18. Specifically, the shape correction unit 124 does not reduce the shape of the object OB in all parts within the blind spot area, but reduces the shape of only the part where a contact force is applied from the fingertip to the object OB. The part where a contact force is applied from the fingertip to the object OB is the above-mentioned grip center primitive P Bis set. As a result, for example, in FIG. 16 , the shape of the object OB is reduced in the X-axis direction, but the shape of the object OB is not reduced in the Y-axis direction. As a result, even if a measurement error is included in the X-axis direction, the "prismatic two-finger" grasping posture can be maintained while satisfying the constraint that the contact force of each finger is parallel to the X-axis. Furthermore, if the shape of the object OB is reduced in the Y-axis direction, the number of fingers that can grasp the object OB decreases (the grasping area becomes smaller), which may result in the need to change the grasping posture. However, in this embodiment, the shape of the object OB is not reduced in the Y-axis direction, but is reduced only in the X-axis direction where the contact force is applied, so that the same grasping posture can be maintained while allowing for measurement error.

[0068] Returning to the description of the flowchart in Fig. 9, the command generation unit 125 then generates a control command for controlling each actuator 12 of the robot 10 based on the gripping posture determined by the gripping posture determination unit 123, the shape of the object OB partially reduced by the shape correction unit 124, and the state data acquired by the acquisition unit 121 (step S110).

[0069] For example, the command generation unit 125 determines the control amount of each actuator 12 required to take a gripping posture for the object OB, the shape of which has been partially reduced, while still satisfying the constraint conditions, and generates a control command including the control amount.

[0070] Next, the communication control unit 126 transmits a control command to the robot 10 via the communication interface 110 (step S112). Upon receiving the control command, the operation control unit 14 of the robot 10 controls the actuator 12 based on the control command. This causes the robot 10 to act and execute the task of grasping the object OB. This ends the processing of this flowchart.

[0071] According to the embodiment described above, the control device 100 of the robot 10 measures the shape of the object OB based on an image of the object OB captured by the camera 11 and estimates the center of the object OB. The control device 100 selects one gripping posture as the gripping posture of the object OB from among multiple gripping postures in the gripping taxonomy based on the center of the object OB and the center of a reference object OBref, whose relationship with the gripping posture is defined in the gripping taxonomy. When causing the robot 10 to grip the object OB in the selected gripping posture, the control device 100 reduces a portion of the shape of the object OB that may contain measurement errors. Specifically, the control device 100 reduces a portion that is in the blind spot of the camera 11 and where a contact force is applied from the fingertips to the object OB when the object OB is gripped in the selected gripping posture. This makes it possible to use a multi-degree-of-freedom hand that can assume multiple grasping postures, making it possible to grasp an object OB more robustly against measurement errors and more stably, even if the object OB has an unknown shape or a complex shape.

[0072] The above-described embodiment can be expressed as follows. a storage medium for storing computer-readable instructions; a processor connected to the storage medium; The processor executes the computer-readable instructions to: determining a center of a first object to be grasped by the end effector for each of a plurality of possible grasping positions of the end effector; measuring a shape of the second object based on an image of the second object captured from at least one viewpoint; selecting one gripping posture when causing the end effector to grip the second object from among the plurality of gripping postures based on the center of the second object whose shape has been measured and the center of the first object; When the end effector is caused to grasp the second object in the selected grasping posture, a part of the measured shape of the second object is reduced, the part being in a blind spot when viewed from the viewpoint and through which a force is applied from the end effector to the second object. Control device.

[0073] The above-described embodiment can also be expressed as follows. measuring a shape of the object to be grasped based on an image of the object to be grasped captured from at least one viewpoint; a step of selecting one of the plurality of gripping postures when the end effector grips the gripping object, based on a center of a reference object associated with each gripping posture that the end effector can take, the center of the reference object that is expected to be gripped by the end effector, and a center of the gripping object whose shape has been measured; a step of reducing a part of the measured shape of the object to be grasped, which part is in a blind spot when viewed from the viewpoint, and through which a force is applied from the end effector to the object to be grasped, when the end effector is caused to grasp the object to be grasped in the selected grasping posture; An object grasping method comprising:

[0074] The above describes the form for carrying out the present invention using an embodiment, but the present invention is not limited to such an embodiment, and various modifications and substitutions can be made within the scope that does not deviate from the gist of the present invention. [Explanation of symbols]

[0075] 10...robot, 11...camera, 12...actuator, 13...status sensor, 14...motion control unit, 100...control device, 110...communication interface, 120...processing unit, 121...acquisition unit, 122...shape measurement unit, 123...grasping posture determination unit, 124...shape correction unit, 125...command generation unit, 126...communication control unit, 130...storage unit

Claims

1. determining a center of a first object to be grasped by the end effector for each of a plurality of possible grasping positions of the end effector; measuring a shape of the second object based on an image of the second object captured from at least one viewpoint; selecting one gripping posture when the end effector grips the second object from among the plurality of gripping postures based on the center of the second object whose shape has been measured and the center of the first object; a step of reducing only a part of the measured shape of the second object, which part is in a blind spot when viewed from the viewpoint and to which a force is applied from the end effector to the second object, when the end effector is caused to grasp the second object in the selected grasping posture; An object grasping method comprising:

2. The step of selecting the gripping posture includes: estimating a center of the second object whose shape has been measured; and selecting, from the plurality of gripping postures, a gripping posture associated with the first object having the same center as the estimated center of the second object. The object gripping method according to claim 1 .

3. The plurality of gripping positions includes a first gripping position in which a side surface of a horizontal cylinder is gripped. The object gripping method according to claim 1 or 2.

4. the first gripping posture includes a constraint condition that directs a force applied from the end effector to the side surface of the horizontal cylinder when the end effector comes into contact with the side surface toward a central axis, which is the longitudinal direction of the horizontal cylinder; The object gripping method according to claim 3 .

5. The plurality of gripping postures includes a second gripping posture for gripping a sphere. The object gripping method according to claim 1 .

6. the second gripping posture includes a constraint condition that directs a force applied from the end effector to the surface of the sphere when the end effector contacts the surface of the sphere toward a center point of the sphere. The object gripping method according to claim 5 .

7. The plurality of gripping positions includes a third gripping position in which an end of a vertical cylinder is gripped. The object gripping method according to claim 1 .

8. The third gripping posture includes a constraint condition that directs a force applied from the end effector to the end portion of the vertical cylinder when the end effector comes into contact with the end portion toward a central axis, which is a longitudinal direction of the vertical cylinder. The object gripping method according to claim 7.

9. The plurality of gripping postures include a fourth gripping posture in which an end of a rectangular parallelepiped is gripped. The object gripping method according to any one of claims 1 to 8.

10. The fourth gripping posture includes a constraint condition that makes a force applied from the end effector to the end portion of the rectangular parallelepiped when the end effector comes into contact with the end portion parallel to a central axis, which is a longitudinal direction of the rectangular parallelepiped. The object gripping method according to claim 9.

11. On the computer, determining a center of a first object to be grasped by the end effector for each of a plurality of possible grasping positions of the end effector; measuring a shape of the second object based on an image of the second object captured from at least one viewpoint; selecting one gripping posture when the end effector grips the second object from among the plurality of gripping postures based on the center of the second object whose shape has been measured and the center of the first object; a step of reducing only a part of the measured shape of the second object, which part is in a blind spot when viewed from the viewpoint and to which a force is applied from the end effector to the second object, when the end effector is caused to grasp the second object in the selected grasping posture; A program to execute.

Images