JP2014097572A - Gripping robot and method for gripping - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for gripping an object reliably and completely at a low calculation cost.SOLUTION: A gripping robot comprises: a gripping section comprising an arm section, a pair of hand sections provided in the arm section, and finger sections provided in each of the pair of hand sections, for gripping an object on a workbench; a camera section provided in the gripping section for capturing an image of the object; and a controlling section for controlling the gripping section and the camera section. The controlling section: specifies a direction of gravitational force in a captured image captured by the camera section; detects, on the basis of the direction of gravitational force in the captured image, a boundary section between the object and the workbench in the captured image; and controls the gripping section on the basis of a position of the boundary section so as to insert each finger section of the pair of hand sections into between the workbench and the object.

Description

  The present invention relates to a gripping robot and a gripping method.

  Conventionally, a method of gripping an object (work) by a robot is to place a sensor unit such as a force sensor or a tactile sensor on the fingertip or wrist of the robot, and grip the object based on output information from the sensor unit. (For example, refer to Patent Document 1).

JP 2010-5732 A

However, the conventional gripping method includes, as described in Patent Document 1, the frictional force between the object and the fingertip, the weight of the object, the moving speed of the hand, etc. in addition to the output information from various sensor units. It is necessary to solve complex equations of motion that take into account. That is, in general, the conventional gripping method requires complicated control such as impedance control, position control, and force control, and thus has a problem that calculation cost (calculation resource, calculation time) is enormous.
In addition, since the calculation time is long, there is a risk that the object (particularly a heavy object) may fall if the object is uncertainly grasped due to a delay in grasping control.

  In addition, the conventional gripping mode is a method of gripping the side surface of the object, but there is a concern that the object cannot be safely gripped because it is necessary to grip the side surface with a strong force that does not cause the object to slide down. is there. For example, a slippery and soft object such as a cylindrical object made of aluminum may be deformed or scratched by gripping the side surface.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a technique for securely and completely grasping an object with low calculation cost.

In order to solve the above problem, a gripping robot according to an aspect of the present invention includes an arm unit, a pair of hand units provided on the arm unit, and finger units provided on each of the pair of hand units. A gripping part that grips an object on a table, a camera part that is provided in the gripping part and images the object, and a control part that controls the gripping part and the camera part, The control unit identifies a gravitational direction in the captured image captured by the camera unit, detects a boundary between the object and the platform in the captured image based on the gravitational direction in the captured image, The gripping portion is controlled so that each finger portion of the pair of hand portions is inserted between the table and the object based on the position of the boundary portion.
According to the above configuration, it is not necessary to solve a complicated equation of motion (the equation of motion taking into account the frictional force between the object and the finger, the weight of the object, the moving speed of the hand portion, etc.). Can be reduced. Therefore, the object can be gripped with a low calculation cost. Furthermore, since the calculation cost is low, there is little possibility of delay in gripping control, and the object can be gripped reliably. In addition, since the gripping mode is such that a finger part is inserted between the base and the object and the object is supported and grasped from below, the object can be reliably grasped. Therefore, for example, the risk of the object falling is reduced. Furthermore, since it is not a gripping mode in which the object is supported and gripped from below and the side surface of the object is gripped with a strong force, the object can be safely gripped.
In other words, according to the above configuration, the complicated calculation using the output information from the force sensor is reduced as much as possible, and the gripping mode for supporting and gripping the object from the bottom is ensured. In addition, the object can be gripped.

In the gripping robot, the control unit may specify a direction of gravity in the captured image based on control information of the gripping unit at the time of capturing the captured image.
According to the above configuration, the direction of gravity can be easily specified based on the information originally provided.

The gripping robot further includes a sensor unit capable of measuring or calculating a direction of gravity in the gripping unit or the camera unit, and the control unit is configured to output the captured image based on output information from the sensor unit. The direction of gravity in the captured image may be specified.
According to the above configuration, the direction of gravity can be specified even when, for example, a gyro sensor or a magnetic sensor is used as a sensor unit that can measure or calculate the direction of gravity in the gripping unit or the camera unit.

In the gripping robot, the control unit inserts each finger portion of the pair of hand portions between the table and the target object based on the position of the boundary portion, and the side surface of the target object. The grip portion may be controlled so as to press each of the pair of hand portions.
According to the said structure, a target object can be hold | gripped more reliably and more completely.

In order to solve the above problem, a gripping method according to another aspect of the present invention includes an arm part, a pair of hand parts provided on the arm part, and finger parts provided on each of the pair of hand parts. A gripping method for gripping an object on a table, means for capturing the object, means for specifying a gravitational direction in a captured image captured by the imaging means, and the captured image Means for detecting a boundary portion between the object and the platform in the captured image based on a gravitational direction in the image, and the pair of the space between the table and the object based on a position of the boundary portion. And means for controlling the gripping means so as to insert each finger part of the hand part.
According to the above configuration, as with the above-described gripping robot, complicated calculation using output information from the force sensor is reduced as much as possible, and the gripping mode in which the object is supported and gripped from below is reduced, so the calculation cost is low The object can be gripped reliably and completely.

1 is a schematic perspective view of a gripping robot according to an embodiment of the present invention. It is an example of the functional block diagram which shows the function of a control part. It is explanatory drawing for demonstrating the holding | grip operation | movement of a holding robot. It is explanatory drawing for demonstrating the holding | grip operation | movement of a holding robot. It is a flowchart which shows an example of operation | movement of a holding robot.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a schematic perspective view of a gripping robot 1 according to an embodiment of the present invention. As shown in FIG. 1, the gripping robot 1 includes a fixing unit 10, an arm unit 20, a hand unit 30, a finger unit 40, and a camera unit 50. The arm part 20, the hand part 30, and the finger part 40 constitute a grip part.

  The fixing unit 10 is fixed on, for example, a floor, a wall, a ceiling, or a movable carriage. The fixing unit 10 includes a control unit 11 (illustrated by a broken line) inside. The control unit 11 controls the grip unit (arm unit 20, hand unit 30, finger unit 40) and camera unit 50. Details of the control unit 11 will be described later.

The arm unit 20 includes a first frame 21, a second frame 22, a third frame 23, a fourth frame 24, and a fifth frame 25.
The first frame 21 is connected to the fixed part 10 via a rotating shaft. The second frame 22 is connected to the first frame 21 and the third frame 23 via a rotation shaft. The third frame 23 is connected to the second frame 22 and the fourth frame 24 via a rotation shaft. The fourth frame 24 is connected to the third frame 23 and the fifth frame 25 via a rotation shaft. The fifth frame 25 is connected to the fourth frame 24 via a rotation shaft. Under the control of the control unit 11, the arm unit 20 moves while the frames 21 to 25 rotate in a complex manner around the rotation axes.

  A pair of hand portions 30 is provided on the fifth frame 25 of the arm portion 20. The hand unit 30 includes a first frame 31, a second frame 32, a third frame 33, and a pressing unit 34. The first frame 31 is connected to the fifth frame 25 of the arm unit 20 via a rotation shaft. The second frame 32 is connected to the first frame 31 and the third frame 33 via a rotation shaft. The third frame 33 is connected to the second frame 32 via a rotating shaft. The pressing part 34 is connected to the tip of the third frame 33. In the hand unit 30, the frames 31 to 33 are complexly rotated around the rotation axes and moved under the control of the control unit 11.

  A finger part 40 is provided on the pressing part 34 of each hand part 30. The finger unit 40 includes a first frame 41, a second frame 42, and a third frame 43. The first frame 41 is connected to the pressing unit 34 of the hand unit 30 via a rotation shaft. The second frame 42 is connected to the first frame 41 and the third frame 43 via a rotation shaft. The third frame 43 is connected to the second frame 42 via a rotation shaft. Under the control of the control unit 11, the fingers 40 move by moving the frames 41 to 43 around the rotation axes in a complex manner. The finger part 40 is a member that directly grips the object.

  A camera unit 50 is provided on the fifth frame 25 of the arm unit 20. The camera unit 50 images the object under the control of the control unit 11.

  Next, the control unit 11 will be described. FIG. 2A is an example of a functional block diagram illustrating functions of the control unit 11. 3 and 4 are explanatory diagrams for explaining the gripping operation of the gripping robot 1. Note that FIG. 2B will be described later.

  As illustrated in FIG. 2A, the control unit 11 includes a camera unit control unit 100, a captured image acquisition unit 110, a gravity direction identification unit 120, a boundary detection unit 130, and a gripping unit control unit 140.

  The camera unit control unit 100 controls imaging by the camera unit 50. The captured image acquisition unit 110 acquires a captured image captured by the camera unit 50. For example, the captured image acquisition unit 110 acquires a captured image P obtained by capturing an object W on the table D as illustrated in FIG.

  The gravity direction identification unit 120 identifies the gravity direction in the captured image acquired by the captured image acquisition unit 110. Specifically, the gravitational direction specifying unit 120 is based on the control information of the gripping unit control unit 140 (information defining the position and orientation of the gripping unit) at the time of capturing the captured image. Is identified. More specifically, the gravity direction specifying unit 120 uses a homogeneous transformation matrix based on the control information of the first frame 21, the second frame 22, the third frame 23, the fourth frame 24, and the fifth frame 25. The position and orientation of the fifth frame 25 provided with the camera unit 50, that is, the position and orientation of the camera unit 50 itself are obtained, and the direction of gravity in the captured image is specified from the position and orientation of the camera unit 50. For example, the captured image acquisition unit 110 specifies the gravity direction G in the captured image P as illustrated in FIG. For convenience of explanation, in FIGS. 3B and 3C, the captured image P is rotated so that the gravity direction P is directly below.

  The boundary detection unit 130 detects the boundary between the object and the base in the captured image based on the direction of gravity in the captured image specified by the captured image acquisition unit 110. For example, the boundary detection unit 130 detects a broken line portion C, which is a boundary between the object W and the table D, as shown in FIG. In addition, although each dotted line part shown in FIG.3 (c) is a boundary part of the target object W and a background, the boundary part detection part 130 is the boundary part of the target object W and the stand D from the gravity direction G. Is not detected.

  The gripping unit control unit 140 controls the gripping units (arm unit 20, hand unit 30, finger unit 40). For example, the gripping unit control unit 140 moves the gripping unit to move the camera unit 50 to the imaging position of the object.

  In addition, the gripping unit control unit 140 moves the gripping unit to grip the object. Specifically, the gripping unit control unit 140 moves each finger unit 40 of the pair of hand units 30 between the platform and the object based on the position of the boundary unit detected by the boundary unit detection unit 130. Control the gripping part to insert. More specifically, as shown in FIG. 3 (c), the gripper control unit 140 is a point on a broken line C that is a boundary between the object W and the table D, and a point A (in the drawing). A point B (point indicated by a triangle in the figure) immediately below a point indicated by a quadrangle is extracted as a gripping point. The point A is a reference point for extracting the gripping point B, and is a point obtained from a mark or pattern registered in advance or a feature amount obtained by image processing. The gripping unit control unit 140 that has extracted the gripping point B controls the gripping unit so that the finger unit 40 is inserted into the gripping point B.

  Thereby, the grasping robot 1 supports the object W from below, for example, as shown in FIG. In the example shown in FIG. 4A, the gripping robot 1 supports the bottom surface of the object W with the second frame 42 and the third frame 43 of the finger unit 40. In FIG. 4, a cross-sectional view of the object is used for convenience of explanation. For example, the surface S in FIGS. 4A and 4B is a cross section obtained by cutting the object W along a cut surface (a surface indicated by a one-dot chain line) passing through the reference point A and the gripping point B in FIG. is there. FIG. 4C shows a conventional gripping mode shown for comparison. In the conventional gripping mode, as shown in FIG. 4 (c), the side of the object W is gripped with a strong force that does not cause the object W to slide down, and therefore the object W may be deformed. However, in the gripping mode shown in FIG. 4A, most of the load on the weight is applied to the second frame 42 and the third frame 43 that support the bottom surface, so that the force from the side is in contact with the object W. There is no fear that the object W is deformed.

  In addition, the gripping unit control unit 140 inserts each finger unit 40 of the pair of hand units 30 between the base and the object based on the position of the boundary detected by the boundary detection unit 130. The grip portion may be controlled so as to press each of the pair of hand portions 30 against the side surface of the object.

  Thereby, for example, as shown in FIG. 4 (b), the gripping robot 1 grips the object W while gently pressing the side surface while supporting the object W from below. In the example shown in FIG. 4B, the gripping robot 1 presses the side surface of the object W against the hand unit 30 while supporting the bottom surface of the object W with the second frame 42 and the third frame 43 of the finger unit 40. It is supported by the part 34. Note that the point supported by the pressing unit 34 may be a reference point A on the side surface as shown in FIG. 4B, for example.

  Next, the operation of the gripping robot 1 will be described using a flowchart. FIG. 5 is a flowchart showing an example of the operation of the gripping robot 1. Note that it is assumed that the gripping robot 1 holds information (for example, a registration mark, a template image, and the like) related to a reference point of the target object for extracting a gripping point of the target object.

  The gripping unit control unit 140 moves the camera unit 50 to the imaging position where the object is imaged (step S10). The camera part control part 100 makes the camera part 50 image an object. The captured image acquisition unit 110 acquires a captured image captured by the camera unit 50 (step S12).

  The gripping unit control unit 140 reads information (for example, a registration mark, a template image, etc.) related to the reference point of the target object, and tries to detect the reference point of the target object by image processing (for example, pattern matching). That is, the reference point is searched (step S14).

  When the reference point is not detected by the grip control unit 140 (step S16: No), the object is imaged from another imaging position, and the reference point is detected again (step S10 to step S16).

  When the reference point is detected by the grip control unit 140 (step S16: Yes), the gravity direction specifying unit 120 calculates the position and orientation of the camera unit 50 using the homogeneous transformation matrix (step S18). The gravity direction specifying unit 120 that has calculated the position and orientation of the camera unit 50 specifies the gravity direction in the captured image (step S20).

  The boundary detection unit 130 detects the boundary between the object and the base in the captured image based on the gravity direction in the specified captured image (step S22). The gripping unit control unit 140 attempts to extract a gripping point that is a point on the boundary and immediately below the reference point, that is, searches for the gripping point (step S24).

  When a gripping point is not extracted by the gripping part control unit 140 (step S26: No), the object is imaged from another imaging position, and the gripping point is extracted again (steps S10 to S26).

  When the gripping point is extracted by the gripping unit control unit 140 (step S16: Yes), the gripping unit control unit 140 moves the gripping unit so that the finger unit is inserted into the bottom surface of the object and the arm unit is attached to the side surface. (Step S28), and the gripping unit is controlled to lift the object (Step S30). Then, this flowchart ends.

  As described above, according to the gripping robot according to the embodiment of the present invention, the complicated motion equation (the motion equation considering the frictional force between the object and the finger part, the weight of the object, the moving speed of the hand part, etc.) Since there is no need to solve, calculation cost can be reduced. Therefore, the object can be gripped with a low calculation cost. Furthermore, since the calculation cost is low, there is little possibility of delay in gripping control, and the object can be gripped reliably. In addition, since the gripping mode is such that a finger part is inserted between the base and the object and the object is supported and grasped from below, the object can be reliably grasped. Therefore, for example, the risk of the object falling is reduced. Furthermore, since it is not a gripping mode in which the object is supported and gripped from below and the side surface of the object is gripped with a strong force, the object can be safely gripped. In other words, according to the above configuration, the complicated calculation using the output information from the force sensor is reduced as much as possible, and the gripping mode in which the object is supported and gripped from the bottom is ensured. In addition, the object can be gripped.

  In the above-described embodiment, as shown in FIG. 4 (a), the gripping mode in which the bottom surface of the object is supported by the finger part 40, and the bottom surface of the object is supported by the finger part 40 as shown in FIG. 4 (b). Although the grip mode in which the side surface of the object is pressed by the hand unit 30 has been described, the grip mode is not limited to this and may be other grip modes. For example, it is good also as a holding | grip aspect which hold | suppresses the side surface of a target object with another finger part, supporting the bottom face of a target object with a certain finger part.

  In the above-described embodiment, the method of specifying the gravitational direction based on the control information of the gripping unit at the time of capturing the captured image has been described as the method of identifying the gravitational direction in the captured image. The method is not limited to this and may be another method. For example, the gravity direction in the captured image may be specified based on output information from a sensor unit that can measure or calculate the direction of gravity.

  When the gravitational direction in the captured image is specified based on the output information from the sensor unit, the gripping robot 1 can measure or calculate the direction of gravity in the gripping unit or the camera unit 50 (for example, a gyro sensor or a magnetic sensor). ). The fixing unit 10 of the gripping robot 1 includes a control unit 12 that is different from the control unit 11 inside. FIG. 2B is an example of a functional block diagram illustrating functions of the control unit 12.

  2B, the gravity method specifying unit 122 specifies the direction of gravity in the captured image based on output information from a sensor unit (for example, a gyro sensor or a magnetic sensor) at the time of capturing the captured image. That is, instead of step S18 and step S20 in the flowchart shown in FIG. 5, the gravity direction specifying unit 122 specifies the gravity direction in the captured image based on the output information from the sensor unit (eg, gyro sensor, magnetic sensor). To do. In FIG. 2B, members denoted by the same reference numerals as the members in FIG. 2A are the same as the members having the same reference numerals in FIG.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes designs and the like that do not depart from the gist of the present invention.

  For example, the number of frames of the arm unit 20, the hand unit 30, and the finger unit 40 is not limited to the number shown in FIG. The camera unit 50 may be provided not on the arm unit 20 but on the hand unit 30 and the finger unit 40.

  DESCRIPTION OF SYMBOLS 1 ... Gripping robot 10 ... Fixed part 11, 12 ... Control part 20 ... Arm part 30 ... Hand part 40 ... Finger part 50 ... Camera part 100 ... Camera part control part 110 ... Captured image acquisition part 120, 122 ... Gravity direction specific | specification part 130: boundary detection unit 140: gripping unit control unit

Claims (5)

A gripping part that has an arm part, a pair of hand parts provided on the arm part, and a finger part provided on each of the pair of hand parts;
A camera unit that is provided in the gripping unit and images the object;
A control unit for controlling the gripping unit and the camera unit,
The controller is
Identify the direction of gravity in the captured image captured by the camera unit,
A boundary portion between the object and the platform is detected in the captured image based on a gravitational direction in the captured image, and the pair between the platform and the object is detected based on a position of the boundary portion. A gripping robot that controls the gripping part to insert each finger part of the hand part. The controller is
The gripping robot according to claim 1, wherein a gravitational direction in the captured image is specified based on control information of the gripping unit at the time of capturing the captured image. The gripping robot is
A sensor unit capable of measuring or calculating the direction of gravity in the gripping unit or the camera unit;
The controller is
The gripping robot according to claim 1, wherein a gravitational direction in the captured image is specified based on output information from the sensor unit at the time of capturing the captured image.   The control unit inserts each finger part of the pair of hand parts between the table and the object based on the position of the boundary part, and the pair of hand parts on a side surface of the object. The gripping robot according to any one of claims 1 to 3, wherein the gripping unit is controlled to press each of the hand units. A gripping method for gripping an object on a table using an arm part, a pair of hand parts provided on the arm part, and a finger part provided on each of the pair of hand parts,
Means for imaging the object;
Means for specifying the direction of gravity in the captured image captured by the imaging means;
Means for detecting a boundary portion between the object and the base in the captured image based on a direction of gravity in the captured image; and between the base and the object based on a position of the boundary portion. And a means for controlling the gripping means so as to insert each finger part of the pair of hand parts.

Images