JP4001105B2 - Grasping method of arbitrarily shaped object by robot - Google Patents

Description

  The present invention relates to a method of gripping an object having an arbitrary shape by a robot, and more particularly to a method of gripping an object by recognizing the shape of an object using a visual sensor.

In an industrial robot or the like, usually, the object to be grasped and the processing content are limited, and the shape of the robot hand and the grasping control according to the object are performed. Furthermore, there is a robot hand using a visual sensor as a technique for enlarging a gripping object from a single item to a plurality of items and performing processing corresponding to each (see, for example, Patent Documents 1 and 2).
The technique of Patent Document 1 aims to improve the operation speed of the hand by making the finger center axis of the hand coincide with the optical axis of the optical distance sensor and the visual sensor so that the coordinate conversion calculation is unnecessary. The technique of Patent Document 2 is to place an imaging means for imaging a workpiece, which is an object, on a robot hand, calculate the position and shape of the workpiece based on the captured image, and control the gripping posture of the hand. .
JP-A-6-31666 JP 2003-94367 A

  These techniques are based on the premise that the object to be grasped has a grasping portion that is assumed to be grasped by the robot hand, or has a form suitable for grasping by the robot hand. However, when the application range of the robot hand is further expanded, it is necessary to perform an operation of appropriately grasping various objects having arbitrary shapes.

  Therefore, an object of the present invention is to provide a method for gripping an arbitrarily shaped object by a robot that has a visual sensor and can appropriately grasp an arbitrarily shaped object by a robot hand.

In order to solve the above problems, a method for gripping an arbitrarily shaped object by a robot according to the present invention is a method for grasping an arbitrarily shaped object by a robot having a visual sensor and a robot hand having multiple fingers and multiple joints, 1) From the image of the target object acquired by the visual sensor, a simple shape that approximates the target object shape is obtained from image recognition among several predetermined simple shapes, and (2) the obtained simple shape and its orientation and size are determined. based on, the cross-section of a target object including a gripping point section of the object including rectangular gripping and gripping point is holding pattern as a rectangle to select the appropriate holding method and a circular gripping a grip form a circular Set the number of fingers used for gripping, the gripping direction, the gravity center position of the gripping point, and the placement of the gripping point, and (3) calculate the gripping point of each finger of the robot hand accordingly, and (4) Control The gripping operation is performed.

When it is determined that the predetermined simple shape includes a rectangular parallelepiped and the target object approximates a rectangular parallelepiped, grip points are arranged on the two surfaces having the smallest distance between the two surfaces, which do not include the floor surface. The gripping operation may be performed by square gripping.

If an object of rectangular approximation gripping the square grip has a central grip plane the gripping point in the rectangular parallelepiped are arranged, it may match the center of gravity of the gripping point on the gripping plane.

The predetermined simple shape comprises an oval sphere, if the target object is determined to approximate the ellipsoid, gripping operation by a circular gripping placing gripping point on intersection line between a horizontal plane and the surface of the object passing through the center of the object It is good to do.

  When gripping an object that is an ellipsoidal approximation by circular gripping, among the fingers involved in gripping, the position of the center of gravity of the gripping point of the finger placed closer to the body of the robot than the target object, and the robot from the target object Each gripping point may be arranged so that the line connecting the center of gravity of the gripping point of the finger placed on the side far from the body passes through the approximate center of the elliptic sphere.

Furthermore, when it is determined that the predetermined simple shape includes a cylinder and the target object approximates a cylinder, the gripping point is arranged on the intersection of the cylinder surface and the plane including the center axis depending on the direction of the center axis of the cylinder. The gripping operation may be performed by selecting either the square gripping to be performed or the circular gripping in which the gripping point is arranged on the end surface of the cylinder.

  When gripping an object with a cylindrical approximation by square grip, among the fingers involved in gripping, the center of gravity position of the grip point of the finger placed closer to the robot body than the target object, and the robot body from the target object It is preferable to arrange each gripping point so that a line connecting the center of gravity of the gripping point of the finger placed on the far side intersects the approximate center axis of the cylinder.

  In the present invention, first, an image of an object is picked up by a visual sensor, and image recognition is performed from the acquired image, so that the object is applied (approximated) to a simple shape (a rectangular parallelepiped, a cylinder, an elliptic sphere, etc.). . An appropriate gripping method is selected from the approximate shape, its orientation, and size from among square gripping and circular gripping, and a gripping point by the selected gripping method is calculated. The movement path of each finger to the gripping point is obtained by calculating backward from the obtained gripping point, each finger is controlled according to the obtained movement path, and the object is gripped.

It is determined whether or not the finger of the robot hand can be placed at the calculated gripping point. If it is determined that the finger cannot be placed, the gripping surface or the gripping point may be rotated about a predetermined axis and retried.

Depending on the positional relationship between the object and the robot, the finger may not be placed so that the set correspondence between the gripping point and the finger is satisfied. In such a case, the gripping surface or the gripping point is rotated around a predetermined axis , and a retry is performed, that is, the gripping point is set and the movement path of each finger to the gripping point is set.

  According to the present invention, it is possible to grip an arbitrary object having a relatively complicated shape by fitting an object to a simple shape and setting a gripping method and a gripping point according to the fitted shape. The application range of robot hands is expanded. In addition, since the fitting method and the program for selecting the gripping point can be made common and simplified by performing the fitting, the program capacity required for the control can be reduced, and reliable gripping can be performed.

Even if it is difficult to grip at the initially set gripping point due to the positional relationship between the object and the robot, the gripping point adjustment can be simplified by rotating the gripping surface and gripping point around a predetermined axis. It is possible to suppress gripping and prevent the object from being gripped.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings. In order to facilitate the understanding of the description, the same reference numerals are given to the same components in the drawings as much as possible, and duplicate descriptions are omitted.

  FIG. 1 is a diagram showing a configuration of a robot hand to which a method of gripping an arbitrarily shaped object by a robot according to the present invention is applied. FIG. 2 is a diagram illustrating a control system thereof, that is, gripping an arbitrarily shaped object by a robot according to the present invention. It is a block block diagram which shows the control system which performs a method.

  The robot hand 1 according to the present embodiment includes four fingers: a thumb 10, an index finger 11, a middle finger 12, and a ring finger 13. The thumb 10 is a four-degree-of-freedom link system in which a motor 14 is arranged for each of four joints, and a six-axis force sensor 15 is arranged at the root. The six-axis force sensor 15 is a sensor that can independently detect a force in a direction along each of the three axial directions orthogonal to each other and a force around each axial direction. The other three fingers 11 to 13 are configured such that the motor 14 is arranged for each of the three joints excluding the most advanced joint, and the most advanced joint is configured to interlock with an adjacent joint (interlocking joint 17). Therefore, each of the four joints constitutes a link system having three degrees of freedom. The degree of freedom of the entire robot hand 1 is 13. A 6-axis force sensor 15 is arranged at the base of each finger 11 to 13 like the thumb 10. The tips 10t to 13t of these fingers are composed of curved elastic bodies. Further, an encoder potentiometer 16 for detecting the bending angle of the joint is disposed at each joint.

  Although not shown in FIGS. 1 and 2, the robot hand 1 is attached to the robot body 8 by an arm 7. The robot body 8 has arms 7 on the left and right sides and a camera eye 3 on the head. The arm 7 is also provided with a motor and an encoder potentiometer so that the robot hand 1 can be moved to a desired position.

  The control system is mainly configured by a control ECU 2 including a RAM, a ROM, a CPU, and the like, and recognizes the shape and position of an object by image recognition from an output image of a camera eye 3 that captures an image of the grasped object. An image recognizing unit 20, a gripping posture calculating unit 21 that calculates a gripping position / griping posture based on the recognition result, and a hand control unit 22 that controls the movement of the robot hand 1. The 6-axis force sensor 15 and the output signals of the encoder potentiometer 16 are input to the hand control unit 22, and the movement of each motor 14 is controlled by the motor driver 4.

  The control system is not limited to such a configuration, and the control unit that controls the movement of the robot hand 1 so as to have the instructed posture, and the image recognition device and the gripping posture calculation unit may be separated. . Further, it may be divided by hardware or software.

  Next, a control method for gripping a gripping object by the robot hand 1 will be specifically described. FIG. 3 is a main flowchart of the control process.

  First, based on the image information input from the camera eye 3, the image recognition unit 20 applies the gripping object to one of the predetermined simple shapes by image recognition (step S1). Here, it is assumed that fitting (approximation) to any simple shape among a rectangular parallelepiped (including a cube), a cylinder, and an elliptical sphere (including a perfect sphere) is performed. In this fitting, a minimum simple shape including the grasped object is obtained. 4 to 6 show examples of fitting in the case where an arbitrarily shaped object, a cup, and an apple are fitted to a rectangular parallelepiped, a cylinder, and an elliptic sphere, respectively. This fitting can be performed by template matching, feature extraction, correlation calculation, or the like. If it is applied to the simple shape, the size and orientation of the applied simple shape are obtained (step S3).

  In step S5, the type of the applied simple shape is determined, and the process branches to steps S11, S31, and S41 according to the determined simple shape.

  In step S11, a gripping method determination process is performed when the fitted simple shape is a rectangular parallelepiped. A flowchart of specific processing is shown in FIG. First, the gripping method is set to square gripping (step S110). Here, the square grip is a grip form in which the cross section of the object including the grip point is rectangular. Then, the two surfaces having the shortest distance between the two surfaces in the combination of the two surfaces facing each other (there are three sets) are set as the gripping plane (step S112).

  Next, it is determined whether one of the two surfaces is in contact with the floor surface (step S114). If it is in contact with the floor surface, the combination of the two opposing surfaces (two sets) that has a shorter distance between the two surfaces is set as the gripping plane (step S116). If it is not in contact with the floor, the selected grip plane is maintained.

  Here, when there are two or more pairs of opposing two planes having the shortest distance between the two surfaces (when either cross section is a square), the combination including the plane closer to the robot body 8 is given priority. If the distances are the same, a combination including a plane on the outer side with respect to the robot body 8 is given priority.

When gripping plane is determined and branches to determine the main axis of the length L _M of the object (step S118). Here, the main axis in the rectangular parallelepiped refers to a straight line that passes through the center of the object and is parallel to the long side direction constituting the grip plane. When the grip plane is a square, a straight line parallel to the side direction closer to the robot body 8 is used as the main axis.

When the length L _M in the main axis direction is short and less than L ₀ , the process proceeds to step S120 to hold two fingers, the length L _M in the main axis direction is sufficiently long, and L ₁ (L ₁ > L ₀ ), The process moves to step S122 to hold four fingers, and if the length L _M in the main axis direction is L ₀ or more and L ₁ or less, the process goes to step S124 and is set to hold three fingers. To do.

  When the number of fingers used for gripping is determined, the process proceeds to step S126 to determine the spindle direction. When it is determined that the main axis direction is horizontal (for example, when the angle formed with the horizontal plane is less than 45 degrees), the gripping direction is set from above (step S128). If it is determined that the main axis direction is not horizontal but close to vertical, the gripping direction is set from the horizontal direction (step S130).

  When the gripping direction, the gripping surface, and the number of fingers used for gripping are determined, the process proceeds to step S13 to obtain the spatial coordinate position of the gripping point.

  FIG. 8 shows gripping point positions in the case of four-finger gripping. Here, it represents with the rectangular parallelepiped 60 to which the target object 6 was fitted, The holding plane 60A, 60B and the main axis | shaft 60C are set to this rectangular parallelepiped 60. FIG. The center 60O of the rectangular parallelepiped 60 is set on the main axis 60C, and the intersections of the holding planes 60A and 60B and the straight line 60D orthogonal to the main axis 60C and the holding plane 60A (the plane closer to the robot of the both holding planes 60A and 60B). A gripping point 10X by the thumb 10 is set, and an intersection of the straight line 60D and the gripping plane 60B is a gripping point 12X by the middle finger 12. Then, two points (the center of the two points are 12X) that are separated by 12 from 12X in the direction of the main axis 60C on the gripping plane 60B are defined as a gripping point 11X by the index finger 11 and a gripping point 13X by the ring finger 13, respectively.

  In the case of gripping with two fingers, as shown in FIG. 9, grip points 10X and 12X are obtained by the same method as in the case of gripping with four fingers. In addition, you may hold | grip with the index finger 11 instead of the middle finger 12 (In this case, the holding point 11X will be arrange | positioned in the holding point 12X position).

  In the case of a three-finger instruction, the basic arrangement is the same as in the case of four-finger gripping. The gripping point position is shown in FIG. Also in this case, intersections between the straight lines 60D orthogonal to the gripping planes 60A and 60B and the main axis 60C and the gripping planes 60A and 60B are obtained. The point at which the intersection point with 60A is the gripping point 10X by the thumb 10 is the same as in the case of gripping two fingers or four fingers, but the intersection point of the straight line 60D and the gripping plane 60B is set as the gripping point of any finger. Instead, this intersection is set as the midpoint of the gripping point 11X by the index finger 11 and the gripping point 12X by the middle finger 12. Here, the gripping points 11X and 12X are arranged at a distance of Δl in the direction of the main shaft 60C.

  When the gripping point coordinates are determined, a gripping posture for placing each finger at the gripping point is calculated by an inverse kinematic method (step S15). Specifically, each posture and position are calculated from the tip toward the robot body in the order of each finger, wrist, and arm.

  If the grip posture can be calculated, it is determined whether or not the arm joint angle is within the movable range (step S17). If it is out of the movable range, the process proceeds to step S19 to perform grip point adjustment processing. This gripping point adjustment processing is performed as follows in the case of a rectangular parallelepiped.

  FIG. 11 shows an example of adjustment when the gripping direction is set from the lateral direction. If the initial gripping direction is from the direction indicated by 1 in the figure, this is changed in the order of 1 → 2 → 3 → 4 by 90 degrees counterclockwise as viewed from above. That is, the gripping plane is moved counterclockwise to the adjacent surface.

  An adjustment example when the gripping direction is set from above is shown in FIG. Assume that the initial gripping direction is from the direction indicated by 5 in the figure (where the thumb is placed on the front side of the figure. In other cases, the number indicates the side closer to the thumb) ). Also in this case, the gripping direction is changed by 90 degrees counterclockwise when viewed from above in the order of 5 → 6 → 7 → 8. That is, it is the same as the case from the side in that the gripping plane is moved counterclockwise to the adjacent surface.

  In the next step S <b> 21, it is determined whether or not the gripping direction adjustment limit position has been reached. Specifically, in the case of a rectangular parallelepiped, the case of returning to the initial setting position is determined as the limit position, it is determined that gripping is impossible, the process proceeds to step S23, the gripping is impossible, and the process ends. . If the gripping direction has not reached the initial position, it is determined that the limit position has not been reached, and the process returns to step S13 to perform recalculation from the calculation of the gripping point position coordinates.

  If it is determined in step S17 that the arm is within the movable range, the process proceeds to a gripping operation. Specifically, the target gripping posture is obtained by controlling the motors of the arm and wrist (step S25). Next, the hand control unit 22 controls each motor 14 of the robot hand 1 by the motor driver 4 so that the output of the encoder potentiometer 16 is at a predetermined angular position, thereby obtaining the grip position where the tip of each finger is obtained. (Step S27). Then, gripping is performed by controlling the operation of the motor 14 so that the gripping force calculated from the output of the six-axis force sensor 15 becomes a target value (step S29). After gripping, the program shifts to another program, and a predetermined operation such as movement of the object and posture change is executed.

If the simple shape fitted in step S5 is an elliptical sphere, the process branches to step S31 to perform a gripping method determination process in the case of an elliptical sphere. FIG. 13 shows a flowchart of detailed processing in this case. First, the circular grip from above is set (step S310). Circular gripping is a gripping form in which a cross section of an object including a gripping point is elliptical (including a true circle). First, the number of fingers used for gripping is determined. Specifically, the major axis D _L in the horizontal cross section including the center of the object is compared with the threshold D ₀ (step S312). When the long diameter D _L is less than the threshold D ₀ , the two fingers are gripped (step S 314), and when the long diameter D _L is the threshold D ₀ or more, the four fingers are gripped (step S 316). .

  When the gripping method is determined, the process proceeds to step S13 to determine the gripping point position. A method of obtaining the gripping point position will be described with reference to FIG. Here, the elliptical sphere fitted with the object 6 is defined as 62, the horizontal section including the center 62O is 62D in the elliptical sphere 62, and the circumference that is the intersection of the horizontal section 62D and the surface of the elliptical sphere 62 is defined as the circle. A vertical line vector passing through the center 62O is represented by 62V and 62V.

  The spatial position coordinate of the base of the arm 7 is R1, and the spatial position coordinate at a position that is a predetermined distance h away vertically downward is R2. Here, when the spatial position coordinates of R1 are represented by (x, y, z), the spatial position coordinates of R2 can be represented by (x, y, z-h). A plane including R1, R2, and 62O is obtained, two intersection points of the plane P and the circumference 62R are obtained, and the side closer to the robot body 8 is set as a gripping point 10X by the thumb 10, and the other point is set by the middle finger 12. The gripping point is 12X (see FIG. 15A). In the case of two-finger gripping, the gripping point setting process ends here.

  In the case of four-finger gripping, two points on the circumference 62R that are separated by a distance Δl from the gripping point 12X are set as a gripping point 11X by the index finger 11 and a gripping point 13X by the ring finger 13, respectively (FIG. 15B). reference). Instead of the distance Δl, the gripping points 11X and 13X may be arranged at positions where the gripping point 12X is shifted on the circumference 62R by ± θ with respect to the center 62O.

  When the gripping point coordinates are determined, the gripping posture is calculated (step S15). The subsequent processing is the same as that when the fitted shape is a rectangular parallelepiped except for the gripping point adjustment processing. Therefore, only the gripping point adjustment process with different processes will be described.

  The gripping point adjustment process in step S19 is performed as follows in the case of circular gripping. Each gripping point position shown in FIGS. 15A and 15B is moved by 5 degrees counterclockwise around the vertical axis on the circumference 62R. In step S39, when the movement angle exceeds 180 degrees from the initial setting position, it is determined as the limit position and it is determined that the grip is impossible.

If the shape fitted in step S5 is a cylinder, the process proceeds to step S41 to perform a gripping method determination process. A flowchart of specific processing is shown in FIG. First, the direction of the spindle is determined (step S410). If it is determined that the main axis direction is the horizontal direction (when the angle between the main axis and the vertical direction is a predetermined angle, for example, 45 degrees or more), that is, it is determined that the cylinder is lying down, the process proceeds to step S412. Then, determine the length of the cylinder and branch. When the length L _{P of the} cylinder is short and less than L ₀ , the process proceeds to step S414 to hold two fingers, the length L _{P of the} cylinder is sufficiently long, and L ₁ (L ₁ > L ₀ ) is set. when it exceeds the set to transition to four-finger grip to step S416, if the length L _P of the cylinder is L ₀ or L ₁ or less is set to migrate to three fingers gripping the step S418.

Step S414, in the case of setting the two fingers or three fingers grasping at S418 is further compared with the diameter _{D P} and the threshold Dth of the cylinder (step S420). If D _P > Dth and it is determined that the shape is closer to the disk than the cylinder, the shape is changed to square gripping with four fingers from the lateral direction (step S422). And if the D _P is determined to less Dth, if you set the four-finger grip in step S416, it sets the square gripping from above (step S424).

Principal axis direction at step S410 is when it is determined that the vertical direction, the process proceeds to step S426 length of the cylinder (height) comparing the _{L P} and the threshold Lth (Lth> _{L 1).} If L _P > Lth, the cylinder is determined to be a sufficiently long cylinder, and is set to square gripping with four fingers from the side (step S428). If L _P is less than or equal to Lth, it is determined that the cylinder is short and is set to circular gripping from above (step S430), and the diameter D _{P of} the cylinder and the threshold value Dth2 are compared (step S432).

When it is determined that D _P > Dth2 and the cylinder is sufficiently thick, the four-finger grip is set (step S434). If it is determined that the thin cylindrical D _P is not enough thickness in Dth2 less Conversely, setting the two fingers gripping (step S436).

  When the gripping method is determined, the process proceeds to step S13 to determine the gripping point position. A method for obtaining the gripping point position in this case will be described with reference to FIGS. Here, a cylinder fitted with the object 6 is represented by 61, the main axis of the cylinder 61 is 61c, its center is 61O, the wall surface is 61w, and both end surfaces are represented by 61A and 61B, respectively.

  In the case of rectangular gripping, it is basically similar to that of a rectangular parallelepiped. First, when gripping the end surface of the disk, as shown in FIG. 17A, each end point 61A, 61B is used as a gripping plane, and each gripping point is processed in the same manner as in the case of the rectangular parallelepiped 60 shown in FIG. 10X to 13X are arranged. As a result, the gripping points 10X and 12X are arranged on the central axis.

  When gripping a horizontal cylinder from above (in the case of holding four fingers), as shown in FIG. 17B, a horizontal straight line 61D passing through the center 61O of the cylinder 61 and orthogonal to the main axis 61c is obtained. The intersection of this and the wall surface 61w is defined as a gripping point 10X by the thumb 10 and a gripping point 12X by the middle finger 12, respectively. Then, the gripping point 12X is sandwiched, and the points separated by Δl in the main axis direction are set as the gripping point 11X by the index finger 11 and the gripping point 13X by the ring finger 13, respectively.

  A case where a vertically oriented cylinder is gripped from the lateral direction will be described with reference to FIGS. As in the case shown in FIG. 14, the spatial position coordinate of the base of the arm 7 is R1, and the spatial position coordinate of a position that is a predetermined distance h away vertically downward is R2. The points where R1 and R2 are shifted by the distance s toward the robot body 8 are denoted as R1 'and R2'. As shown in FIG. 19, in the case of gripping a cylinder from the lateral direction, the object 6 is positioned closer to the robot body 8 in the left-right direction in the figure than the base portion of the robot hand 1 and the arm 7 portion. In consideration of this, R1 and R2 are shifted by a predetermined distance toward the robot body 8 side. The shift amount s is appropriately set according to the distance between the fingertip and the extension line at the center of the arm 7 when the object is gripped from the lateral direction.

  A plane P ′ passing through the set R 1 ′, R 2 ′ and the center 61 O of the cylinder 61 is obtained. A cross section 61s passing through the center 61O of the cylinder 61 and orthogonal to the main axis 61c is obtained, and the intersections of the cross section 61s and the plane P 'are defined as a gripping point 10X by the thumb 10 and a gripping point 12X by the middle finger 12, respectively. The gripping point 11X by the index finger 11 and the gripping point 13X by the ring finger 13 can be obtained as points separated from the gripping point 12X by Δl in the main axis direction, respectively.

  In the case of circular grip from above, grip points are arranged on the end faces. Specifically, as in the case of FIG. 18, the plane P ′ is obtained, and the intersections between the plane P ′ and the circumference of the end surface 61 </ b> A are set as a gripping point 10 </ b> X by the thumb 10 and a gripping point 12 </ b> X by the middle finger 12, respectively. The gripping point 11X by the index finger 11 and the gripping point 13X by the ring finger 13 are arranged by the same method as in the case of an elliptic sphere.

  When the gripping point coordinates are determined, the gripping posture is calculated (step S15). The subsequent processing is the same as that when the fitted shape is a rectangular parallelepiped except for the gripping point adjustment processing. The gripping point adjustment process in the case of a cylinder may be moved around the main axis as in the case of an elliptical sphere when the cylinder is vertically oriented. Even when the cylinder is in the horizontal direction, it is moved around the main axis. In this case, the tip or the main body of the robot hand 1 may collide with the floor surface. Therefore, the angle that becomes the limit position is changed according to the diameter of the cylinder.

  By arranging the gripping points according to the shape in this way, it is possible to place the gripping points on a line or a plane including the object center as much as possible. By setting the gripping point so as to surround the center of the object that is close to the center of gravity, it is possible to set a gripping point that is highly likely to grip the target object reliably. When a thick and short cylindrical object with a similar shape is placed vertically, the end face is gripped, but in this case as well, the end face itself is close to the center of gravity, so the target object is securely gripped. There is no change to the point where a gripping point that is likely to be set can be set.

  And by applying a simple shape to any object with a relatively complex shape, the number of cases where the shape cannot be recognized and cannot be gripped is reduced, and a suitable gripping point candidate for an object of any shape Can be set. In addition, since the method of setting the gripping point is simplified, the size of the program for setting can be reduced and the setting can be performed at high speed, so that the responsiveness of the robot hand is improved.

  Furthermore, even if there is a slight difference between the fitted shape and the actual shape, the gripping force can be absorbed by controlling the gripping force so that it becomes the target value, so the accuracy of the fitting is improved. It is not necessary to have high accuracy, and the object can be gripped appropriately.

  Here, the method of determining the gripping point after applying the object to any one of a rectangular parallelepiped, an elliptical sphere, and a cylinder has been described, but the simple shape to apply is not limited to these three, and these three are It is not essential.

  Also, the number of fingers is not limited to four, and may be two, three, or five or more if there are two or more opposing fingers. Further, the combination of opposing fingers is not limited to the one-finger-finger combination as described above, and may be a combination of two fingers-2 fingers, two fingers-3 fingers, or the like.

  The method for setting the gripping point is not limited to the above-described processing routine, and may be set as appropriate according to the arrangement of the fingers of the robot hand and the movable range based on square gripping and circular gripping.

It is a figure which shows the structure of the robot hand to which the holding method of the arbitrarily-shaped object by the robot which concerns on this invention is applied. It is a block block diagram which shows the control system which performs the holding | grip method of the arbitrarily-shaped object by the robot which concerns on this invention. It is a main flowchart of the control processing of the holding | grip method which concerns on this invention. It is a figure which shows the example which applied the arbitrarily-shaped object to the rectangular parallelepiped. It is a figure which shows the example which applied the cup to the cylinder. It is a figure which shows the example which applied the apple to the elliptical sphere. It is a flowchart which shows the holding | grip method determination processing of a rectangular parallelepiped. It is a figure which shows the holding | grip point position in the case of holding a rectangular parallelepiped with four fingers. It is a figure which shows the holding | grip point position in the case of holding a rectangular parallelepiped with two fingers. It is a figure which shows the holding | grip point position in the case of holding a rectangular parallelepiped with three fingers. It is a figure explaining adjustment of a grasping position when grasping a rectangular parallelepiped from a horizontal direction. It is a figure explaining adjustment of a grasping position when grasping a rectangular parallelepiped from an upper direction. It is a flowchart which shows the grasping method determination process of an ellipsoid. It is a figure explaining the setting of the grip point position of an ellipsoid. It is a figure which shows the holding | grip point position when holding an elliptical sphere with two fingers (FIG. 15A) and holding with four fingers (FIG. 15B). It is a flowchart which shows the holding | grip method determination process of a cylinder. It is a figure which shows the holding | grip point position in the case of carrying out square holding of the cylinder. It is a figure explaining the setting of the holding point position of a vertically oriented cylinder. It is a figure which shows the holding | grip point position of the vertical cylinder from a horizontal direction. It is a figure which shows the holding | grip point position of the vertical cylinder from upper direction.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Robot hand, 2 ... Control ECU, 3 ... Camera eye, 4 ... Motor driver, 6 ... Object, 7 ... Arm, 8 ... Robot main body, 10 ... Mother finger, 11 ... Indicating finger, 12 ... Middle finger, 13 ... Ring finger 10t to 13t ... tip, 10X to 13X ... each gripping point, 14 ... motor, 15 ... axial force sensor, 16 ... encoder potentiometer, 17 ... interlocking joint, 20 ... image recognition unit, 21 ... gripping posture calculation unit, 22 ... hand control unit, 60 ... cuboid, 61 ... column, 62 ... elliptical sphere.

Claims (8)

A method of gripping an arbitrarily shaped object by a robot having a visual sensor and a robot hand having multiple fingers and multiple joints,
A simple shape that approximates the target object shape among several predetermined simple shapes is obtained by image recognition from the image of the target object acquired by the visual sensor, and the gripping point is determined based on the obtained simple shape and its orientation and size. the cross section of the target object with the cross section of the object including rectangular gripping and gripping point is holding pattern as a rectangle to select the appropriate holding method and a circular gripping a grip form a circular containing, used in the gripping fingers The gripping point, the gripping direction, the gravity center position of the gripping point, and the placement of the gripping point, the gripping point of each finger of the robot hand is calculated accordingly, and the gripping operation is performed by controlling each finger. A method of gripping an arbitrarily shaped object by a featured robot. When it is determined that the predetermined simple shape includes a rectangular parallelepiped and the target object approximates a rectangular parallelepiped, the gripping point is given priority on the two surfaces with the smallest distance between the two surfaces that do not include the floor surface. 2. The method for gripping an arbitrarily shaped object by a robot according to claim 1, wherein the gripping operation is performed by square gripping arranged in a vertical direction. When gripping the rectangular gripping the object rectangular approximation, claim 2, wherein the center of the gripping plane the gripping point in the rectangular parallelepiped are arranged, to match the center of gravity of the gripping point on said gripping plane A method for gripping an arbitrarily shaped object by the robot described above. When it is determined that the predetermined simple shape includes an elliptical sphere and the target object approximates an elliptical sphere, a gripping operation is performed by circular gripping in which a gripping point is arranged on the intersection line between the horizontal plane passing through the center of the object and the surface of the object. The method for gripping an arbitrarily shaped object by the robot according to claim 1, wherein:   When gripping an object that is an ellipsoidal approximation by circular gripping, among the fingers involved in gripping, the position of the center of gravity of the gripping point of the finger placed closer to the body of the robot than the target object, and the robot from the target object 5. An arbitrarily shaped object by a robot according to claim 4, wherein each grip point is arranged so that a line connecting the center of gravity of the grip point of the finger placed on the side far from the body passes through the approximate center of an elliptical sphere. Gripping method. When it is determined that the predetermined simple shape includes a cylinder and the target object approximates to a cylinder, a gripping point is arranged on the intersection line between the surface of the cylinder and a plane including the center axis according to the direction of the center axis of the cylinder. 2. The method for gripping an arbitrarily shaped object by a robot according to claim 1 , wherein a gripping operation is performed by selecting either a square grip or a circular grip in which a grip point is arranged on an end face of a cylinder.   When gripping an object with a cylindrical approximation by square grip, among the fingers involved in gripping, the center of gravity position of the grip point of the finger placed closer to the robot body than the target object, and the robot body from the target object 7. The robot according to claim 6, wherein each of the grip points is arranged so that a line connecting the center of gravity of the grip point of the finger placed on the far side intersects the approximate center axis of the cylinder. Grasping method. It is determined whether or not the finger of the robot hand can be placed at the calculated gripping point, and when it is determined that the finger cannot be placed, the gripping surface or the gripping point is rotated around a predetermined axis and retried. A method for gripping an arbitrarily shaped object by the robot according to claim 1.

Images