WO2012066819A1 - ワーク取り出し装置 - Google Patents
ワーク取り出し装置 Download PDFInfo
- Publication number
- WO2012066819A1 WO2012066819A1 PCT/JP2011/065961 JP2011065961W WO2012066819A1 WO 2012066819 A1 WO2012066819 A1 WO 2012066819A1 JP 2011065961 W JP2011065961 W JP 2011065961W WO 2012066819 A1 WO2012066819 A1 WO 2012066819A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- workpiece
- hand
- gripping
- work
- posture
- Prior art date
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1612—Programme controls characterised by the hand, wrist, grip control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J19/00—Accessories fitted to manipulators, e.g. for monitoring, for viewing; Safety devices combined with or specially adapted for use in connection with manipulators
- B25J19/02—Sensing devices
- B25J19/04—Viewing devices
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J13/00—Controls for manipulators
- B25J13/08—Controls for manipulators by means of sensing devices, e.g. viewing or touching devices
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/39—Robotics, robotics to robotics hand
- G05B2219/39543—Recognize object and plan hand shapes in grasping movements
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/40—Robotics, robotics mapping to robotics vision
- G05B2219/40053—Pick 3-D object from pile of objects
Definitions
- the present invention relates to a work picking device that picks up a work stacked in bulk.
- the workpiece picking device is a device that picks workpieces one by one with a robot hand from a target on which a plurality of workpieces are stacked.
- a conventional workpiece picking device for example, there is a device disclosed in Patent Document 1.
- a partial shape unit into which the workpiece is divided and a priority order corresponding to the partial shape are stored in advance as data, and a plurality of stacked workpieces are stored.
- image processing a plurality of partial shapes that are gripping position candidates are calculated, and the workpiece to be picked and its gripping portion are determined from the calculated candidates in consideration of the above priorities. It was.
- the present invention has been made in view of the above, and an object of the present invention is to provide a workpiece take-out apparatus that can take out a workpiece in a mode in which the amount of data to be held in advance is small and the calculation time is short. To do.
- a workpiece picking device includes a hand that grips a workpiece, a robot that moves the hand to a desired gripping position and posture, and a sensor that three-dimensionally measures the workpiece to obtain workpiece measurement data.
- a storage medium that stores at least hand shape data, an information processing unit that calculates a gripping position and orientation based on data from the sensor and the storage medium, and a gripping position and orientation calculated by the information processing unit.
- a control unit that controls the robot, and the information processing unit includes an optimum gripping candidate creation unit that directly derives a gripping position and orientation based on the workpiece measurement data and the hand shape data.
- the workpiece take-out apparatus of the present invention it is possible to take out a workpiece with the same algorithm even if the workpiece shape changes in a mode in which the amount of data held in advance is small and the calculation time is also short.
- FIG. 6 is an apparatus configuration diagram showing a workpiece removal apparatus according to a second embodiment.
- FIG. 6 is an apparatus configuration diagram illustrating a workpiece take-out apparatus according to a third embodiment.
- FIG. 6 is an apparatus configuration diagram showing a workpiece take-out apparatus according to a fourth embodiment. It is a figure which shows a part of workpiece taking-out apparatus by Embodiment 5.
- FIG. It is a figure which shows the restriction
- FIG. 10 is a diagram illustrating a two-dimensional hand model used in the fifth embodiment. It is a figure showing the processing image of the optimal holding
- FIG. 10 is an apparatus configuration diagram showing a workpiece take-out apparatus according to a fifth embodiment. It is a graph showing the result of evaluating the success rate at the time of taking out a workpiece according to Embodiment 5 for taking out a workpiece that has been difficult to recognize by the method based on the conventional workpiece shape and has not reached the gripping test. is there.
- FIG. 1 is an apparatus configuration diagram showing a workpiece take-out apparatus according to Embodiment 1 of the present invention.
- the workpiece picking device includes at least a recording medium 1, a sensor 2, an information processing unit 3, a control unit 4, a robot 5, and a hand 6.
- the recording medium 1 stores at least hand shape data.
- the sensor 2 acquires the three-dimensional measurement data of the stacked workpieces.
- the information processing unit 3 outputs the grip position / posture of the workpiece based on data from the storage medium 1 and the sensor 2.
- the control unit 4 controls the operation of the robot 5 and further the operation of the hand 6 based on the grip position / posture data obtained by the information processing unit 3.
- the robot 5 makes the hand 6 an arbitrary position and posture based on a command from the control unit 4, and the hand 6 grips a workpiece.
- the sensor 2 may be mounted at the hand of the robot 5 or may be installed at an appropriate fixing destination separately from the robot 5.
- the sensor 2 is, for example, a binocular or multi-lens stereo camera, an active stereo camera including a light projecting unit and a camera such as a laser or a projector, a device using a time-of-flight method, and a robot operation with a single eye. Any factor analysis method, Structure-from-Motion, Structure-and-Motion equipment, motion stereo camera, visual volume intersection method, or the like can be used.
- the hand 6 may be a pinching type, a suction type, or a type that is inserted into a hole and expanded (hereinafter referred to as a holding type).
- the present invention is characterized in that the gripping position / orientation is dynamically determined from the hand shape and measurement data, rather than determining the gripping position / orientation in advance based on the workpiece shape.
- the unit 3 includes an optimum gripping candidate creation unit 30 that dynamically determines where on the workpiece measurement data from the sensor 2 is to be gripped.
- the optimum gripping candidate creation unit 30 includes a grippable feature extraction unit 301, a hand matching unit 302, and a gripping posture candidate selection unit 303.
- the grippable feature extraction unit 301 extracts grippable features from the workpiece measurement data of the sensor 2.
- the hand matching unit 302 matches the hand shape data stored in the recording medium 1 with the features extracted by the grippable feature extraction unit 301 and dynamically creates a plurality of gripping position / posture candidates.
- the gripping posture candidate selection unit 303 selects the gripping position / posture most easily gripped from the gripping position / posture candidates created by the hand matching unit 302.
- the grippable feature extraction unit 301 extracts features that can be gripped by the hand 6 to be used regardless of the overall shape of the workpiece. For example, in the case of the sandwiching type hand 6, a protruding portion or an edge that is easy to pick up corresponds. In the case of a suction type hand, a surface of a certain area or more corresponds. Further, in the case of an extension type hand, a circular hole generated on the surface of the work corresponds.
- These features can be extracted by applying edges, surfaces, circles, etc. to the 3D measurement data.
- a mode for obtaining these features at high speed for example, a mode in which edge detection using a Canny operator or a Sobel operator is performed on a distance image in which distance is expressed by brightness, circle extraction by Hough transform is performed. Examples include a mode in which a surface is extracted by labeling a region surrounded by edges, a mode in which a texture portion is detected as a protrusion using the edge gradient direction and intensity of luminance, and the like. These are required for basic image processing, and can be completed very quickly even if many features are extracted.
- the feature extracted by the grippable feature extracting unit 301 and the hand shape data (model) stored in the recording medium 1 are matched.
- This is realized by having a model as shown in FIG. 2 according to each hand type.
- the clamping type it can be defined by the opening width of the hand arm immediately before the gripping approach, the approach depth, the vertical width and the horizontal width of the hand.
- the suction type it can be defined by the radius of suction
- the extension type it can be defined by the radius of the hole that can be inserted, the radius of the work surface around the hole, and the penetration depth. Since the hand shape data as described above has simple parameters, the amount of data required is small.
- it may be calculated from CAD data of the hand 6 or the hand 6 may be directly measured.
- the hand shape model For matching with the hand shape model as described above, for example, if it is a suction type, a part with a high matching between the model and the feature point group is a candidate, the feature surface is small, or a hole is opened in the feature surface. When the degree of coincidence with the model is low, the matching score is lowered from the candidate.
- Such matching is realized by matching points such as ICP (Iterative Closest Point) if it is a 3D model, by template matching if it is a 2D model, or by matching by convolution processing considering the model as a filter. Can do.
- interference with the surrounding environment can be taken into consideration at the same time in the case of the sandwiching type or the extension type.
- the hand matching unit 302 calculates a gripping position / posture having the highest matching score for each of the features extracted by the grippable feature extracting unit 301, and uses the plurality of gripping position / postures as gripping position / posture candidates. Or, for a single feature extracted by the grippable feature extraction unit 301, a plurality of gripping position / postures whose matching score is higher than a predetermined threshold are calculated, and the plurality of gripping position / postures are calculated as gripping positions. Let it be a posture candidate. By such processing, the gripping position / posture can be dynamically defined even if the workpiece shape is unknown.
- the score is obtained by two-dimensional image matching with three degrees of freedom of translation two degrees of freedom and one degree of freedom of rotation. It is possible to calculate, and it is possible to calculate a gripping position / posture that is easy to grasp without interference of each feature at high speed.
- the gripping posture candidate selection unit 303 selects the gripping position / posture candidate that is most easily gripped from the gripping position / posture candidates created by the hand matching unit 302. For this, the hand matching unit 302 having the highest score may be selected, or an additional priority may be given. For example, the workpiece at the highest position in the stacking scene may be selected by using the average height and the center of gravity position of the features extracted by the grippable feature extraction unit 301.
- the control unit 4 controls the operation of the robot 5 and further the operation of the hand 6 based on the gripping position candidate data, The hand 6 grips and takes out a target workpiece from a plurality of workpieces stacked in bulk.
- the workpiece take-out device configured as described above, if at least hand shape data and a hand shape model based on the hand shape data are held as prior information, the workpiece shape and the unloading scene state are obtained. Regardless, it is possible to dynamically calculate the gripping position and posture that is easy to grip with less interference between the peripheral workpiece and the hand, and the shape of the workpiece has changed in a manner that requires less data to be held in advance and less calculation time.
- the workpiece can be taken out with the same algorithm.
- the processing of having to redefine the priorities of the parts in the work each time the hand shape is different is not compulsory, it is possible to solve the problem that the labor and calculation time associated with changing the priorities increases. it can.
- Embodiment 2 the hand shape data is directly matched with the workpiece measurement data, that is, the optimum gripping position / posture is calculated using only the hand shape data.
- the workpiece shape data is further used, and the entanglement state between the workpieces that cannot be determined only by the measurement data, or whether the pickup operation is successful. It is also possible to estimate such.
- work shape data is stored in the storage medium 1 in addition to the hand shape data, and as shown in FIG.
- a workpiece state determination unit 31 that evaluates the gripping position / orientation calculated by 30 based on workpiece shape data is further included.
- the workpiece state determination unit 31 includes a workpiece matching unit 311, a workpiece entanglement state estimation unit 312, a workpiece pickup operation estimation unit 313, and a gripping posture candidate selection unit 314.
- the workpiece entanglement state estimation unit 312 determines the entanglement between the workpieces in the bulk stack that does not appear in the workpiece measurement data by determining the interference between workpiece models created from the matching of workpiece measurement data and workpiece shape data. Thus, an operation that prioritizes the gripping position / posture corresponding to an object with little entanglement is realized.
- the workpiece picking-up motion estimation unit 313 calculates the position of the center of gravity of the workpiece when gripped from the hand shape data, the workpiece shape data, and the calculated gripping position / posture, and based on the calculated position of the workpiece after gripping An operation that gives priority to a gripping position / posture with a low possibility of the workpiece falling or the tilting of the workpiece after gripping is realized.
- the optimum gripping candidate creation unit 30 in the second embodiment is basically the same as the optimum gripping candidate creation unit 30 in the first embodiment, but has one gripping position / posture that can be evaluated as optimum. It is different not only to decide but also to select a plurality. Further, in that case, a score for prioritizing a plurality of selected gripping position / postures may be created.
- the workpiece matching data is matched with the workpiece shape data by the workpiece matching unit 311 using the gripping position / posture newly calculated by the optimum gripping candidate creation unit 30 and the extracted feature used therefor as an initial candidate.
- the fact that an initial candidate has already been obtained corresponds to the fact that an initial search for work model matching has been made.
- the feature is a set of three-dimensional points or edge points, accurate matching is performed by the ICP method or the like.
- template collation with three degrees of freedom of translation two degrees of freedom and rotation one degree of freedom, collation using a hash, matching by silhouette, and collation by geometric relationship of features may be used. Further, it may be determined in advance which part of the work is a feature that is likely to be extracted by the optimum gripping candidate creation unit 30, and the matching search range may be limited to increase the speed.
- the workpiece entanglement state estimation unit 312 obtains a corresponding workpiece model by matching workpiece measurement data and workpiece shape data with respect to a workpiece corresponding to a certain gripping position and posture and a workpiece group around the workpiece. By analyzing the state of the workpiece model, interference between workpieces is determined. As a result, when a part of each work model exists at the same position and interferes, the gripping position / posture corresponding to the work model is excluded from the candidates, or candidate selection in the gripping posture candidate selection unit 314 is performed. Processes such as lowering the priority. Also, even if there is no interference between the work models at this time, if the work model pick-up operation is simulated and if it collides with other peripheral work models, the corresponding grip position / posture should be excluded from the candidates as well Priorities may be reduced.
- the workpiece picking-up motion estimation unit 313 performs gripping simulation using a certain gripping position / posture and a corresponding work model and a hand shape model corresponding to the gripping posture / posture.
- the position of the center of gravity of the workpiece is calculated, and the gripping position / posture with a high risk of dropping during the picking-up operation is excluded or the priority is reduced.
- the risk of falling may be determined using the Euclidean distance between the hand gripping position prepared as shown in FIG. 2 and the calculated gravity center position of the workpiece as an evaluation index.
- the gripping posture candidate selection unit 314 selects an optimal gripping position / posture based on the priorities evaluated by the workpiece entanglement state estimation unit 312 and the workpiece pick-up operation estimation unit 313. At this time, the score for candidate creation created by the optimum gripping candidate creation unit 30 may be used.
- Each evaluation index may be a linear combination such as summation or a non-linear combination such as integration.
- the work entanglement state estimation unit 312 and the workpiece pick-up operation are mainly based on the score of the optimum gripping candidate creation unit 30.
- a selection method such as rejecting candidates whose priority of the estimation unit 313 is below a certain level may be taken.
- the workpiece take-out device as in the first embodiment, selection of candidates that can be gripped with high accuracy with less hand interference while reducing the amount of data, model registration effort, and calculation time. can do.
- a gripping position / posture in which a workpiece can be taken out is less likely to be entangled between workpieces and to be less likely to fail during a pick-up operation.
- a take-out operation can also be realized.
- Embodiment 3 FIG. Based on FIG. 4, a workpiece takeout device according to the third embodiment will be described.
- data relating to the gripping position / posture of the next work operation is further stored in the recording medium 1 in the second embodiment, and the information processing unit 3 is suitable for the next work as shown in FIG.
- the next operation motion estimation unit for estimating the gripping position candidate is provided.
- the next work motion estimation unit 32 estimates a grip position candidate based on the grip position / posture of the next work motion stored in the recording medium 1. For example, if the next work after the work removal is the assembly of the work to the product being assembled, the gripping position and posture of the work required during the assembly are suitable for the assembly operation. It is limited to that. For example, it is often impossible to assemble a workpiece gripped inside out as it is, and it is difficult to greatly change the posture that the robot takes while gripping. Although minute gripping errors can be tolerated, if gripping with a large error or gripping with a workpiece posture different from the assembly posture is performed, it is necessary to change the workpiece, which increases work time and labor. .
- work can illustrate not only an assembly
- the method for determining whether or not it is necessary to change the position from the gripping position / posture at the time of removal is to simulate the gripping state of the work using the work shape model, the hand shape model, and the gripping position / posture.
- the next work motion estimation unit 32 holds the work shape model held by the hand shape model by the hand shape model at the work position obtained by simulating the next work such as assembly. Match with the workpiece shape model in the ground state. If the robot can calculate the transformation matrix between the positions and orientations of the respective hand shapes in a state where the positions and orientations of the respective workpiece shape models match, it is determined that the operation without changing can be performed as the next operation. If the calculation is impossible, it can be determined that it is necessary to hold it again, such as holding it again or handing it to another robot hand.
- the workpiece can be taken out in a mode in which the amount of data to be held in advance is small and the calculation time is also short, as in the above-described embodiment. In addition to being able to do this, it is also possible to reduce the tact time by reducing the changeover work.
- Embodiment 4 FIG.
- a plurality of gripping position / posture candidates calculated by the optimum gripping candidate creation unit 30 are selected by the work state determination unit 31, and further selected by the next work motion estimation unit 32. This process was adopted.
- a plurality of gripping position / posture candidates calculated by the optimum gripping candidate creation unit 30 are respectively set as a workpiece state determination unit 31 and a next work motion estimation unit 32. The evaluation is performed in parallel, and the evaluation is finally performed by the gripping posture candidate selection unit 33 in a comprehensive manner.
- the candidate score calculated by the optimum gripping candidate creation unit 30, the evaluation values of the work state determination unit 31 and the next work motion estimation unit 32 are linearly combined such as addition or nonlinearly combined such as integration. It may be expressed and evaluated.
- processing that intervenes a selection method such as rejecting candidates whose evaluation values of the work state determination unit 31 and the next work motion estimation unit 32 are below a certain value, mainly based on the score of the appropriate gripping candidate creation unit 30 It may be taken.
- FIG. 6 shows a part of a workpiece take-out apparatus according to Embodiment 5 of the present invention.
- a sandwich-type hand 6 and a sensor 2 that performs three-dimensional measurement are mounted on the tip of the robot 5.
- the robot 5 of FIG. 6 is an articulated type with 6 degrees of freedom, but may be an articulated type with 7 degrees of freedom.
- the robot 5 may be a double-arm type, vertical scalar type, or parallel link type robot.
- the sensor 2 is attached to the tip of the robot 5 and can be operated together. Alternatively, the sensor 2 may be fixedly installed separately from the robot 5 or may be installed on another movable stage.
- ⁇ Restrict the robot and hand movements as shown in Fig. 7 when unloading workpieces. It is assumed that the robot 5 has a coordinate system having X, Y, and Z directions as shown in FIG. At this time, as shown in FIG. 7A, the bulk object in the supply box 7 is operated in the sky above the supply box 7 with two degrees of freedom in translation in the X and Y axes and one degree of freedom in rotation around the Z axis. Align the position and posture of the hand 6 so that Next, as shown in FIG. 7 (b), the workpiece enters the supply box 7 with a motion of one degree of translation in the Z direction, and the work 6 is gripped by the hand 6. Further, as shown in FIG. 7C, the hand 6 is pulled up while holding the workpiece by an operation with one degree of freedom of translation in the Z direction. This is realized by an operation with a total of 4 degrees of freedom including 3 degrees of freedom of translation and 1 degree of freedom of rotation.
- This operation has many advantages. For example, if the movement of 6 degrees of freedom is outside the movement range of the robot 5 and the robot 5 tries to move the movement of 6 degrees of freedom according to the posture of the workpiece, the robot 5 stops due to an error. If it can be realized by the above-described operation with a total of 4 degrees of freedom, there is a high possibility that such a situation can be avoided. Further, it is possible to avoid the danger that the robot 5 and the hand 6 collide with the supply box 7 and the sensor 2 due to the complicated movement of 6 degrees of freedom. It is easy to design a robot motion range, a safety range, and an interference avoidance range with four degrees of freedom. Furthermore, if the movement has a total of four degrees of freedom, the robot 5 can be a low-cost but high-speed vertical scalar type.
- the line-of-sight direction of the sensor 2 is matched with the direction of the approaching motion as shown in FIG.
- the visual line direction of the sensor here indicates the optical axis direction of the camera lens.
- FIG. 9 An example of the distance image is shown in FIG.
- the amount of reflection of light on the surface of the object is contained in each pixel as a luminance value.
- the height of the object in the pixel is within each pixel.
- FIG. 9 it is displayed brighter if it is close to the imaged sensor 2 and darker if it is far away.
- the distance image has an advantage that the amount of processing calculation is lighter than that of the three-dimensional position data because three-dimensional information including the height can be handled as an image.
- many image processing methods that have been used in production sites can be applied.
- the distance image can be obtained by, for example, a three-dimensional measurement method that can obtain height information such as a spatial coding method or a stereo method, and a three-dimensional sensor that realizes the height information.
- FIG. 10 shows a state in which a certain workpiece is piled up and measured with a three-dimensional sensor to obtain a camera image and a distance image.
- the movement of the robot 5 is limited to four degrees of freedom, the line-of-sight direction of the sensor 2 is made coincident with the approaching movement direction, and the distance grasped workpiece distance image is used to select the optimum gripping candidate for picking up the workpiece.
- the following explanation is focused on picking up and picking up, and it does not matter in what posture the work is picked up.
- CAD data, point cloud data, distance data, two-dimensional image data, and partial information work area specific areas, edge lengths, texture features, geometrical relationships of characteristic parts such as holes
- a two-dimensional hand model as shown in FIG. 11 is used. This is a further simplification of the sandwich-type hand shape 2D model shown in FIG. 2, and the tip portion of the hand that hits the workpiece at the moment when the hand 6 enters and grips is represented by a circular interference region. Since there are two interference areas at the tip of the hand for gripping, this two-dimensional hand model can be defined by only two parameters in total, the radius of the circle and the opening width of the hand that determines the positional relationship between the two circles.
- the tip of the hand is a rectangular parallelepiped, it may be expressed by three parameters using length and width instead of the radius of the circle, or may be defined by a parameter representing a polygon or an ellipse.
- the radius of the approximate circle is set large enough to cover the entire interference area, the hand will be able to use the approximate circle as a model to avoid interference. To function correctly.
- Fig. 12 shows a processing image for creating an optimal gripping candidate using a distance image and a two-dimensional hand model.
- FIG. 13 shows a processing flow.
- the segment is extracted from the distance image.
- the segment here refers to a plane or curved surface portion of the work surrounded by the shape edges.
- Segmentation means extracting this segment. This extraction is realized by edge detection from a distance image using a Canny operator or a Sobel operator.
- the edge in the distance image represents the shape edge of the object itself.
- the area delimited by the detection edge corresponds to the plane or curved surface portion of the workpiece. Segments can be extracted by this method even for thin screws and springs, where model matching with the workpiece shape is difficult.
- a candidate selection method for example, a plurality of candidates may be selected by giving priority to the segment at the highest position, or may be selected based on the area of the segment instead of the height information. From these candidates, the gripping position / posture with the highest score (gripability likelihood) by matching described below is output.
- the segment selected as a candidate is matched with the 2D hand model.
- a position that can be stably grasped without colliding is searched. This is realized by searching for a gripping position and posture in which 1) the area of the segment existing within the opening width of the hand is large, and 2) there is no interference between the interference region at the tip of the hand and the periphery of the segment and no collision occurs.
- 1) is the convolution process between the candidate segment and the open width of the hand
- 2) is the convolution process of the interference area between the candidate segment and the tip of the hand, including the area where it may collide. Can do.
- the area of 2) is subtracted from the area of 1), it becomes “an area that can be grasped and cannot be collided”.
- the gripping position and orientation to be calculated are the orientation and position of the hand model when this maximum likelihood is recorded. This is calculated for a single segment selected as a candidate or for multiple segments. In the case of a single segment, the position and orientation of the hand model that records the maximum likelihood within that segment is selected. In the case of a plurality of segments, the position and orientation of the hand model that records the maximum likelihood of the segment having the largest maximum likelihood in each segment is selected.
- the result image of the gripping position / posture in FIG. 12 shows that a calculation result indicating that it is appropriate to grip the workpiece at a position approximately perpendicular to the edge is output at the center of gravity of the part of the workpiece protruding.
- the number 92 represents the “how easy to grasp” with a maximum of 100 points by normalizing the likelihood of gripping with the area.
- the dark gray circle in the middle represents the gripping position, and the two white circles connected by a straight line represent the interference area at the tip of the hand.
- FIG. 14 shows a more easily understood result obtained by calculating the gripping position / posture for several candidate segments in the same manner.
- the score is high where there is no peripheral interference, and the score is low in the part that is likely to hit when entering. Also, it can be seen that the score of the part orthogonal to the edge of the gripping position is high.
- the opening / closing direction of the hand at the gripping position and the direction of the shape edge of the candidate segment should be perpendicular.
- An example is shown in FIG.
- the edge of the candidate segment is detected, and the edge direction for each local region is calculated.
- the degree of orthogonality increases as the edge direction of the shape edge of the candidate segment that intersects the candidate segment is closer to the opening / closing direction of the hand.
- Both evaluations can be reflected by multiplying the above-mentioned grippable likelihood score or adding weights.
- FIG. 16 is an apparatus configuration diagram showing a workpiece take-out apparatus according to the fifth embodiment of the present invention.
- the apparatus includes a recording medium 1 that stores only a two-dimensional hand model, a sensor 2 that generates a distance image of loosely stacked workpieces, and a gripping position and orientation based on the two-dimensional hand model of the storage medium 1 and the distance image of the sensor 2.
- An information processing unit 3 having an optimum gripping candidate generation unit 30 to be calculated, a control unit 4 that controls the operation based on the gripping position and orientation information, a robot 5 that makes the hand an arbitrary position and orientation based on a command from the control unit, and grips a workpiece Hand 6 is included.
- the processing flow performed in the optimum gripping candidate generation unit 30 is the flow of FIG.
- This processing is implemented with a PC of Core2Duo2G and memory 2G with a calculation time of about 0.1 to 0.2 sec. It can be seen that the processing using three-dimensional information is very fast.
- Multiple removal is a post-process, for example, it can be rejected based on camera measurement results or weight judgment, or placed on a flat surface and separated, so only when it is missed. If it failed, the average success rate for all workpieces was 91.5%. In the present embodiment, it is possible to quickly extract a workpiece having a complicated shape, which has been difficult to remove from the bulk stacking without adjusting each workpiece.
- the workpiece can be taken out in a mode in which the amount of data to be held in advance is small and the calculation time is also short.
- the following advantages are also obtained.
- the line-of-sight direction of the sensor is made coincident with the approach direction when the hand approaches the workpiece in order to take out the stacked workpiece, creation of the optimum gripping candidate can be realized at high speed.
- the robot forms a desired gripping position and posture with a limited total of four degrees of freedom, ie, three degrees of freedom in translation in the X, Y, and Z directions and one degree of freedom of rotation around the approach direction axis for picking up the workpiece.
- the robot operation range, safety range, and interference avoidance range can be easily designed, and the start-up of the apparatus is speeded up.
- the optimum gripping candidate creation unit uses a two-dimensional hand model as hand shape data, uses a distance image as work measurement data, and grips from the grippable likelihood obtained by convolution of the hand shape data and the work measurement data. Calculate the position and orientation. Furthermore, the optimum gripping candidate creation unit extracts a plane or a curved surface based on edge detection from the distance image, and calculates the gripping position and orientation by matching the plane or curved surface with the hand shape data. Therefore, the workpiece can be taken out at a high speed, and adjustments depending on the workpiece shape are not required, and the start-up of the apparatus is also accelerated.
- the optimum gripping candidate creation unit calculates the gripping position and orientation from the orthogonality between the direction of the shape edge of the plane or curved surface extracted based on the edge detection from the distance image and the opening / closing direction of the hand. You can increase the success rate.
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Orthopedic Medicine & Surgery (AREA)
- Human Computer Interaction (AREA)
- Manipulator (AREA)
Abstract
Description
図1は、本発明の実施の形態1によるワーク取出し装置を示す装置構成図である。ワーク取出し装置は、記録媒体1と、センサ2と、情報処理部3と、制御部4と、ロボット5と、ハンド6とを少なくとも備えている。
上記実施の形態1では、ハンド形状データを直接、ワーク計測データにマッチングさせる態様をとっており、すなわちハンド形状データのみで最適な把持位置姿勢を計算していた。これに対して、本実施の形態2は、把持位置姿勢を絞り込んだ後に、さらにワーク形状データを利用して、計測データだけでは判断できないワークどうしの絡み状態や、拾上げ動作が成功するか否かの推定をも行えるようにしたものである。
図4に基づいて、本実施の形態3に係るワーク取出し装置を説明する。本実施の形態3は、上記実施の形態2において、さらに、記録媒体1に次作業動作の把持位置姿勢に関するデータが蓄えられ、図4に示されるように、情報処理部3に次作業に適した把持位置候補を推定する次作業動作推定部が設けられたものである。
図4に示した実施の形態3では、最適把持候補作成部30により計算された複数の把持位置姿勢候補を、ワーク状態判断部31において、選り分けてから、次作業動作推定部32において、さらに選り分けるという処理が採用されていた。これに対し、本実施の形態4は、図5に示されるように、最適把持候補作成部30により計算された複数の把持位置姿勢候補を、ワーク状態判断部31、次作業動作推定部32それぞれでパラレルに評価し、最終的にそれらの評価を把持姿勢候補選択部33において総合的に判断するという処理をとるものである。
処理が高速で、ワーク形状に依存せず調整が容易で、汎用的なバラ積みワーク取り出しを実現する実施形態を示す。図6は本発明の実施の形態5によるワーク取出し装置の一部を示す。ロボット5の先に、挟持型のハンド6と3次元計測をするセンサ2とを搭載する。図6のロボット5は、6自由度の多関節型であるが、7自由度の多関節型でもよい。あるいは、ロボット5は、双腕型、垂直スカラー型、パラレルリンク型のロボットでもよい。また、センサ2は、ロボット5の先端に取付けられ、一緒に稼動できるが、これ以外に、ロボット5と別に固定設置してもよいし、別の可動ステージに設置してもよい。
Claims (11)
- ワークを把持するハンドと、
前記ハンドを所望の把持位置姿勢にするロボットと、
ワークを3次元計測してワーク計測データを得るセンサと、
少なくともハンド形状データを蓄えた記憶媒体と、
前記センサ及び前記記憶媒体からのデータをもとに把持位置姿勢を計算する情報処理部と、
前記情報処理部によって計算された把持位置姿勢に基づいて前記ロボットを制御する制御部とを備え、
前記情報処理部は、前記ワーク計測データと前記ハンド形状データとに基づいて、直接、把持位置姿勢を導き出す最適把持候補作成部を含む
ワーク取り出し装置。 - 前記最適把持候補作成部は、前記ハンド形状データに基づくモデルと、前記ワーク計測データとのマッチングから、把持位置姿勢を直接計算する、請求項1のワーク取り出し装置。
- 前記センサの視線方向は、前記ハンドがバラ積みワークを取り出すためにワークに近づくときの進入方向と一致している、請求項1又は2のワーク取り出し装置。
- 前記ロボットがX,Y,Z方向の並進3自由度と、ワークを取り出す進入方向軸周りの回転1自由度という、合計4自由度の制限された動作で所望の把持位置姿勢を形成する、請求項3記載のワーク取り出し装置。
- 前記最適把持候補作成部は、前記ハンド形状データとして2次元ハンドモデルを使い、前記ワーク計測データとして距離画像を使い、そのハンド形状データとワーク計測データ
との畳みこみで求められる把持可能尤度から把持位置姿勢を計算する、請求項3又は4のワーク取り出し装置。 - 前記最適把持候補作成部は、距離画像からのエッジ検出に基づき平面や曲面を抽出し、その平面や曲面とハンド形状データをマッチングすることで把持位置姿勢を計算する、請求項3乃至5の何れか一項のワーク取り出し装置。
- 前記最適把持候補作成部は、距離画像からのエッジ検出に基づき抽出した平面や曲面の形状エッジの方向と、ハンドの開閉方向との直交度から把持位置姿勢を計算する、請求項6のワーク取り出し装置。
- 前記記憶媒体は、ワーク形状データも蓄えており、
前記情報処理部は、前記最適把持候補作成部により計算された把持位置姿勢を、前記ワーク形状データに基づき評価するワーク状態判断部を含む
請求項1乃至7の何れか一項のワーク取り出し装置。 - 前記ワーク状態判断部は、前記ワーク計測データと前記ワーク形状データとのマッチングから作成されるワークモデル同士の干渉を判定することで、前記ワーク計測データには現れないバラ積み内部のワーク同士の絡みを判定し、絡みが少ない対象に対応した把持位置姿勢を優先するワーク絡み状態推定部を含む、請求項8のワーク取り出し装置。
- 前記ワーク状態判断部は、前記ハンド形状データと、前記ワーク形状データと、計算される把持位置姿勢とから、把持したときのワークの重心位置を計算し、それをもとに把持後のワークの落下または把持後のワークの傾きの可能性が低い把持位置姿勢を優先するワーク拾い出し動作推定部を含む、請求項8又は9のワーク取り出し装置。
- 前記記憶媒体は、ワーク取り出しの次の作業に関するワークの次作業把持位置姿勢のデータをも蓄えており、
前記情報処理部は、前記ロボットによって次作業把持位置姿勢への実現が可能な把持位置姿勢を優先する次作業動作推定部を含む
請求項1乃至10の何れか一項のワーク取り出し装置。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020137012534A KR101453234B1 (ko) | 2010-11-17 | 2011-07-13 | 워크 취출 장치 |
JP2012544126A JP5558585B2 (ja) | 2010-11-17 | 2011-07-13 | ワーク取り出し装置 |
DE112011103794.5T DE112011103794B4 (de) | 2010-11-17 | 2011-07-13 | Aufnehmervorrichtung für Werkstücke |
US13/878,321 US9089966B2 (en) | 2010-11-17 | 2011-07-13 | Workpiece pick-up apparatus |
CN201180055387.5A CN103221188B (zh) | 2010-11-17 | 2011-07-13 | 工件取出装置 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2010257082 | 2010-11-17 | ||
JP2010-257082 | 2010-11-17 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012066819A1 true WO2012066819A1 (ja) | 2012-05-24 |
Family
ID=46083763
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2011/065961 WO2012066819A1 (ja) | 2010-11-17 | 2011-07-13 | ワーク取り出し装置 |
Country Status (6)
Country | Link |
---|---|
US (1) | US9089966B2 (ja) |
JP (1) | JP5558585B2 (ja) |
KR (1) | KR101453234B1 (ja) |
CN (1) | CN103221188B (ja) |
DE (1) | DE112011103794B4 (ja) |
WO (1) | WO2012066819A1 (ja) |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014502564A (ja) * | 2010-12-30 | 2014-02-03 | ゼンロボティクス オイ | 把持位置を決定する方法、コンピュータプログラム及び装置 |
CN103659822A (zh) * | 2012-09-14 | 2014-03-26 | 株式会社安川电机 | 机器人设备 |
EP2781315A1 (en) * | 2013-03-18 | 2014-09-24 | Kabushiki Kaisha Yaskawa Denki | Robot picking system, control device and method of manufacturing a workpiece |
JP2014205209A (ja) * | 2013-04-11 | 2014-10-30 | 三菱電機株式会社 | ロボットシステム、及びロボットシステムの制御方法 |
JP2015009314A (ja) * | 2013-06-28 | 2015-01-19 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
JP2015085475A (ja) * | 2013-10-31 | 2015-05-07 | キヤノン株式会社 | 情報処理装置、情報処理方法 |
JP2015100866A (ja) * | 2013-11-22 | 2015-06-04 | 三菱電機株式会社 | ロボットシミュレーション装置、プログラム、記録媒体及び方法 |
JP2015145055A (ja) * | 2014-02-04 | 2015-08-13 | セイコーエプソン株式会社 | ロボット、ロボットシステム、制御装置および制御方法 |
JP2016132086A (ja) * | 2015-01-22 | 2016-07-25 | 三菱電機株式会社 | ワーク供給装置およびワーク把持姿勢計算方法 |
US9401222B1 (en) | 2015-11-23 | 2016-07-26 | International Business Machines Corporation | Determining categories for memory fail conditions |
JP2016221647A (ja) * | 2015-06-02 | 2016-12-28 | キヤノン株式会社 | 把持制御装置、把持制御方法及びプログラム |
US9694499B2 (en) | 2014-09-16 | 2017-07-04 | Fanuc Corporation | Article pickup apparatus for picking up randomly piled articles |
WO2018123087A1 (ja) * | 2016-12-28 | 2018-07-05 | オムロン株式会社 | シミュレーション装置、ロボット、シミュレーション方法、及びそのプログラム |
JP2018126862A (ja) * | 2018-05-14 | 2018-08-16 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
JP2018144144A (ja) * | 2017-03-03 | 2018-09-20 | 株式会社キーエンス | 画像処理装置、画像処理方法、及びコンピュータプログラム |
JP2019063984A (ja) * | 2017-10-02 | 2019-04-25 | キヤノン株式会社 | 情報処理装置、方法及びロボットシステム |
JP2019126886A (ja) * | 2018-01-25 | 2019-08-01 | 株式会社リコー | 情報処理システム、ピッキング箇所特定方法及びプログラム |
JP2019153605A (ja) * | 2018-02-28 | 2019-09-12 | オムロン株式会社 | 落下ばらつき予測装置、方法、プログラム、及びロボットシステム |
US10603788B2 (en) | 2016-09-05 | 2020-03-31 | Fanuc Corporation | Robot simulation apparatus |
JP2020049651A (ja) * | 2019-12-11 | 2020-04-02 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
JPWO2021001882A1 (ja) * | 2019-07-01 | 2021-01-07 | ||
JPWO2021029064A1 (ja) * | 2019-08-15 | 2021-02-18 | ||
WO2021053750A1 (ja) * | 2019-09-18 | 2021-03-25 | 株式会社Fuji | 作業ロボットおよび作業システム |
JP6878727B1 (ja) * | 2020-12-01 | 2021-06-02 | 株式会社エクサウィザーズ | アノテーション支援装置、方法、及びプログラム |
CN112936257A (zh) * | 2021-01-22 | 2021-06-11 | 熵智科技(深圳)有限公司 | 一种工件抓取方法、装置、计算机设备及存储介质 |
JP2021171885A (ja) * | 2020-04-27 | 2021-11-01 | パナソニックIpマネジメント株式会社 | 保持装置、制御方法、制御プログラム、制御システム、およびエンドエフェクタ |
JP2021185016A (ja) * | 2017-03-06 | 2021-12-09 | 株式会社Fuji | 画像処理装置 |
WO2023013743A1 (ja) * | 2021-08-04 | 2023-02-09 | 京セラ株式会社 | 保持パラメータ推定装置及び保持パラメータ推定方法 |
WO2024023934A1 (ja) * | 2022-07-26 | 2024-02-01 | ファナック株式会社 | ワーク取出し装置、ワーク取出し方法及び制御装置 |
Families Citing this family (66)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5837065B2 (ja) | 2011-06-29 | 2015-12-24 | 三菱電機株式会社 | 部品供給装置 |
JP5494617B2 (ja) * | 2011-10-26 | 2014-05-21 | 株式会社安川電機 | ロボットシステムおよび加工品の製造方法 |
US20130343640A1 (en) * | 2012-06-21 | 2013-12-26 | Rethink Robotics, Inc. | Vision-guided robots and methods of training them |
DE102013013114A1 (de) * | 2012-08-17 | 2014-02-20 | Liebherr-Verzahntechnik Gmbh | Vorrichtung zum automatisierten Entnehmen von in einem Behälter angeordneten Werkstücken |
US9233470B1 (en) * | 2013-03-15 | 2016-01-12 | Industrial Perception, Inc. | Determining a virtual representation of an environment by projecting texture patterns |
JP5698789B2 (ja) | 2013-04-18 | 2015-04-08 | ファナック株式会社 | ワークを搬送するロボットの制御装置 |
GB201309156D0 (en) * | 2013-05-21 | 2013-07-03 | Univ Birmingham | Grasp modelling |
JP5929854B2 (ja) * | 2013-07-31 | 2016-06-08 | 株式会社安川電機 | ロボットシステムおよび被加工物の製造方法 |
JP5788460B2 (ja) | 2013-11-05 | 2015-09-30 | ファナック株式会社 | バラ積みされた物品をロボットで取出す装置及び方法 |
JP2015147256A (ja) | 2014-02-04 | 2015-08-20 | セイコーエプソン株式会社 | ロボット、ロボットシステム、制御装置、及び制御方法 |
FR3020303B1 (fr) * | 2014-04-25 | 2016-07-15 | Sileane | Procede et installation de prehension automatique d'un objet. |
JP6425718B2 (ja) * | 2014-05-20 | 2018-11-21 | 株式会社日立製作所 | 組立教示装置及び方法 |
JP6088563B2 (ja) * | 2015-02-10 | 2017-03-01 | ファナック株式会社 | 位置及び姿勢の変換演算機能を備えたワーク取出しロボットシステム、及びワーク取出し方法 |
JP6632208B2 (ja) * | 2015-03-24 | 2020-01-22 | キヤノン株式会社 | 情報処理装置、情報処理方法、プログラム |
US9486921B1 (en) | 2015-03-26 | 2016-11-08 | Google Inc. | Methods and systems for distributing remote assistance to facilitate robotic object manipulation |
US9802317B1 (en) * | 2015-04-24 | 2017-10-31 | X Development Llc | Methods and systems for remote perception assistance to facilitate robotic object manipulation |
US9889564B2 (en) | 2015-07-08 | 2018-02-13 | Empire Technology Development Llc | Stable grasp point selection for robotic grippers with machine vision and ultrasound beam forming |
JP6339534B2 (ja) * | 2015-07-17 | 2018-06-06 | ファナック株式会社 | 最大で二つのワークを把持するハンドを備えたロボットの制御方法およびロボット制御装置 |
EP3383593B1 (en) * | 2015-12-03 | 2021-04-28 | ABB Schweiz AG | Teaching an industrial robot to pick parts |
US9687983B1 (en) * | 2016-05-11 | 2017-06-27 | X Development Llc | Generating a grasp pose for grasping of an object by a grasping end effector of a robot |
JP2018034242A (ja) | 2016-08-31 | 2018-03-08 | セイコーエプソン株式会社 | ロボット制御装置、ロボット、及びロボットシステム |
JP6646894B2 (ja) * | 2016-12-28 | 2020-02-14 | オムロン株式会社 | 保持可否結果出力装置 |
JP6734402B2 (ja) * | 2017-01-12 | 2020-08-05 | 株式会社Fuji | 作業機 |
US10414043B2 (en) | 2017-01-31 | 2019-09-17 | Fanuc America Corporation | Skew and circular boundary for line tracking and circular tracking |
JP6438512B2 (ja) * | 2017-03-13 | 2018-12-12 | ファナック株式会社 | 機械学習により補正した計測データでワークの取り出しを行うロボットシステム、計測データ処理装置および計測データ処理方法 |
CN108618275B (zh) * | 2017-03-20 | 2021-01-26 | Ykk株式会社 | 拉头的搬送装置及其搬送方法 |
US11003177B2 (en) * | 2017-03-24 | 2021-05-11 | Mitsubishi Electric Corporation | Apparatus and method for generating robot program |
US11173615B2 (en) | 2017-03-30 | 2021-11-16 | Soft Robotics, Inc. | User-assisted robotic control systems |
JP6598814B2 (ja) * | 2017-04-05 | 2019-10-30 | キヤノン株式会社 | 情報処理装置、情報処理方法、プログラム、システム、および物品製造方法 |
DE102017207069A1 (de) * | 2017-04-27 | 2018-10-31 | Robert Bosch Gmbh | Prüfvorrichtung zur optischen Prüfung eines Objektes, Produktionsanlage mit der Prüfvorrichtung und Verfahren zur optischen Prüfung des Objektes mit der Prüfvorrichtung |
JP6487495B2 (ja) * | 2017-05-19 | 2019-03-20 | ファナック株式会社 | ワーク取出しシステム |
US10761542B1 (en) | 2017-07-11 | 2020-09-01 | Waymo Llc | Methods and systems for keeping remote assistance operators alert |
US11065761B2 (en) * | 2017-07-25 | 2021-07-20 | Dematic Corp. | Robotic picking training technique |
JP6608894B2 (ja) * | 2017-09-27 | 2019-11-20 | ファナック株式会社 | ロボットシステム |
JP6902208B2 (ja) | 2017-11-14 | 2021-07-14 | オムロン株式会社 | 把持方法、把持システム及びプログラム |
JP6880457B2 (ja) | 2017-11-14 | 2021-06-02 | オムロン株式会社 | 把持方法、把持システム及びプログラム |
US11504853B2 (en) * | 2017-11-16 | 2022-11-22 | General Electric Company | Robotic system architecture and control processes |
JP6879238B2 (ja) * | 2018-03-13 | 2021-06-02 | オムロン株式会社 | ワークピッキング装置及びワークピッキング方法 |
JP6937260B2 (ja) * | 2018-03-19 | 2021-09-22 | 株式会社東芝 | 把持制御装置、把持システム、およびプログラム |
US11584016B2 (en) * | 2018-04-24 | 2023-02-21 | Fanuc Corporation | Robot controller and system |
US10751882B1 (en) * | 2018-05-14 | 2020-08-25 | Amazon Technologies, Inc. | End effector for autonomous object retrieval |
JP6823008B2 (ja) * | 2018-05-18 | 2021-01-27 | ファナック株式会社 | バラ積みされたワークを取り出すロボットシステムおよびロボットシステムの制御方法 |
US11103994B2 (en) * | 2018-07-02 | 2021-08-31 | Teradyne, Inc. | System and method for natural tasking of one or more robots |
US10369701B1 (en) | 2018-10-30 | 2019-08-06 | Mujin, Inc. | Automated package registration systems, devices, and methods |
KR102650494B1 (ko) | 2018-10-30 | 2024-03-22 | 무진 아이엔씨 | 자동화된 패키지 등록 시스템, 디바이스 및 방법 |
EP3670106B1 (en) * | 2018-12-18 | 2024-05-15 | Siemens Aktiengesellschaft | Task-specific robot grasping system and method |
WO2020141579A1 (ja) * | 2019-01-04 | 2020-07-09 | ソニー株式会社 | 制御装置、制御方法、及びプログラム |
US11185978B2 (en) * | 2019-01-08 | 2021-11-30 | Honda Motor Co., Ltd. | Depth perception modeling for grasping objects |
US11148289B1 (en) | 2019-01-08 | 2021-10-19 | Amazon Technologies, Inc. | Entanglement end effector for autonomous object retrieval |
CN113453850A (zh) * | 2019-02-27 | 2021-09-28 | Abb瑞士股份有限公司 | 用于机器人组装的系统和方法 |
JP7299967B2 (ja) * | 2019-03-04 | 2023-06-28 | 株式会社Fuji | シミュレーションシステム |
DE102019107851B4 (de) * | 2019-03-27 | 2022-06-23 | Franka Emika Gmbh | Robotergreifer sowie Verfahren zum Betrieb eines Robotergreifers |
US11745337B2 (en) | 2019-08-29 | 2023-09-05 | Kabushiki Kaisha Toshiba | Handling device, control device, and computer program product |
US11724401B2 (en) | 2019-11-13 | 2023-08-15 | Nvidia Corporation | Grasp determination for an object in clutter |
DE102020128653B4 (de) | 2019-11-13 | 2024-03-14 | Nvidia Corporation | Greifbestimmung für ein Objekt in Unordnung |
CN114845844B (zh) * | 2019-12-17 | 2023-05-30 | 三菱电机株式会社 | 信息处理装置、工件识别装置及工件取出装置 |
JP7364505B2 (ja) | 2020-03-18 | 2023-10-18 | 株式会社東芝 | ハンドリング装置、制御装置、およびプログラム |
DE102020115628A1 (de) | 2020-06-12 | 2021-12-16 | Schunk Gmbh & Co. Kg Spann- Und Greiftechnik | Sensoreinrichtung für ein Greifsystem, Verfahren zur Erzeugung von optimalen Greifposen zur Ansteuerung einer Greifvorrichtung und zugehöriges Greifsystem |
KR102307047B1 (ko) * | 2020-07-03 | 2021-10-01 | 주식회사 두산 | 로봇의 피킹 우선 순위 선정 장치 및 방법 |
CN112873205A (zh) * | 2021-01-15 | 2021-06-01 | 陕西工业职业技术学院 | 基于双夹具实时切换的工业机器人无序抓取方法 |
JP2022131750A (ja) * | 2021-02-26 | 2022-09-07 | 株式会社東芝 | ハンドリングシステム、搬送システム、制御装置、制御プログラム、およびハンドリング方法 |
DE102021002418B3 (de) | 2021-05-07 | 2022-07-14 | Günther Zimmer | Verfahren zum Erstellen von Greiferablaufprogrammen |
JP2023041200A (ja) * | 2021-09-13 | 2023-03-24 | トヨタ自動車株式会社 | ワーク保持装置、ワーク保持方法、プログラム、及び制御装置 |
DE102021212860B4 (de) | 2021-11-16 | 2024-05-08 | Robert Bosch Gesellschaft mit beschränkter Haftung | Verfahren zum Aufnehmen eines Objekts mittels eines Roboters |
CN118058882A (zh) * | 2022-11-24 | 2024-05-24 | 中国科学院深圳先进技术研究院 | 前臂假肢控制方法、装置、设备和介质 |
DE102023201407A1 (de) | 2023-02-17 | 2024-08-22 | Kuka Deutschland Gmbh | Verfahren und System zur Verbesserung der Grifferreichbarkeit |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH03270890A (ja) * | 1990-03-20 | 1991-12-03 | Canon Inc | 視覚センサを備えたワーク把持制御装置 |
JPH09277184A (ja) * | 1996-04-15 | 1997-10-28 | Ricoh Co Ltd | 画像処理方法 |
JPH11198076A (ja) * | 1998-01-05 | 1999-07-27 | Kubota Corp | 保持位置判別装置及び仕分け装置 |
JP2000263482A (ja) * | 1999-03-17 | 2000-09-26 | Denso Corp | ワークの姿勢探索方法および姿勢探索装置ならびにロボットによるワーク把持方法およびワーク把持装置 |
JP2000304509A (ja) * | 1999-04-21 | 2000-11-02 | Matsushita Electric Works Ltd | 物体特定方法及び装置 |
WO2004106009A1 (ja) * | 2003-06-02 | 2004-12-09 | Matsushita Electric Industrial Co., Ltd. | 物品操作システムおよび方法、並びに物品管理システムおよび方法 |
JP2007319997A (ja) * | 2006-06-01 | 2007-12-13 | Nissan Motor Co Ltd | 教示装置および教示方法 |
JP2010210511A (ja) * | 2009-03-11 | 2010-09-24 | Honda Motor Co Ltd | 対象物の3次元位置・姿勢認識装置及びその方法 |
Family Cites Families (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2261069B (en) * | 1991-10-30 | 1995-11-01 | Nippon Denso Co | High speed picking system for stacked parts |
JPH11114860A (ja) * | 1997-10-20 | 1999-04-27 | Yamaha Motor Co Ltd | 認識方法及びピッキング装置 |
JPH11300670A (ja) | 1998-04-21 | 1999-11-02 | Fanuc Ltd | 物品ピックアップ装置 |
JP3768174B2 (ja) | 2002-07-24 | 2006-04-19 | ファナック株式会社 | ワーク取出し装置 |
JP2004160567A (ja) * | 2002-11-11 | 2004-06-10 | Fanuc Ltd | 物品取出し装置 |
SE524796C2 (sv) | 2002-12-10 | 2004-10-05 | Svensk Industriautomation Ab | Kollisionsskydd |
JP2005111607A (ja) * | 2003-10-07 | 2005-04-28 | Fanuc Ltd | ロボット物流トラッキング装置 |
JP4199264B2 (ja) | 2006-05-29 | 2008-12-17 | ファナック株式会社 | ワーク取り出し装置及び方法 |
JP4238256B2 (ja) * | 2006-06-06 | 2009-03-18 | ファナック株式会社 | ロボットシミュレーション装置 |
JP4235214B2 (ja) * | 2006-07-04 | 2009-03-11 | ファナック株式会社 | ロボットプログラムを作成するための装置、プログラム、記録媒体及び方法 |
US7313464B1 (en) * | 2006-09-05 | 2007-12-25 | Adept Technology Inc. | Bin-picking system for randomly positioned objects |
JP4226623B2 (ja) | 2006-09-29 | 2009-02-18 | ファナック株式会社 | ワーク取り出し装置 |
JP4271232B2 (ja) * | 2006-12-20 | 2009-06-03 | ファナック株式会社 | ロボットのオフラインプログラミングを実行するための装置、方法、プログラム及び記録媒体 |
JP4309439B2 (ja) | 2007-03-30 | 2009-08-05 | ファナック株式会社 | 対象物取出装置 |
JP4962123B2 (ja) | 2007-04-27 | 2012-06-27 | 日産自動車株式会社 | 把持候補位置選出装置、把持候補位置選出方法、把持経路生成装置、および把持経路生成方法 |
JP4794011B2 (ja) * | 2008-04-03 | 2011-10-12 | 関東自動車工業株式会社 | 画像処理装置、およびロボット制御システム |
CN102186638B (zh) * | 2008-08-29 | 2016-03-02 | Abb研究有限公司 | 在工业机器人的手臂末端处的工具的顺应装置 |
US8559699B2 (en) * | 2008-10-10 | 2013-10-15 | Roboticvisiontech Llc | Methods and apparatus to facilitate operations in image based systems |
JP5265296B2 (ja) | 2008-10-10 | 2013-08-14 | 本田技研工業株式会社 | ワーク取り出し方法 |
JP2010115723A (ja) * | 2008-11-11 | 2010-05-27 | Seiko Epson Corp | ロボット及びロボットシステム |
JP5304469B2 (ja) | 2009-06-19 | 2013-10-02 | 株式会社デンソーウェーブ | ビンピッキングシステム |
JP5333344B2 (ja) * | 2009-06-19 | 2013-11-06 | 株式会社安川電機 | 形状検出装置及びロボットシステム |
JP5402697B2 (ja) * | 2009-10-26 | 2014-01-29 | 株式会社安川電機 | ロボット装置及びワーク取り出しシステム並びにワーク取り出し方法 |
JP5257335B2 (ja) * | 2009-11-24 | 2013-08-07 | オムロン株式会社 | 3次元視覚センサにおける計測有効領域の表示方法および3次元視覚センサ |
JP5316580B2 (ja) * | 2011-05-17 | 2013-10-16 | 株式会社安川電機 | ロボットシステム |
JP5266377B2 (ja) * | 2011-12-19 | 2013-08-21 | ファナック株式会社 | 物品の姿勢を修正する機能を備えた取出し装置 |
-
2011
- 2011-07-13 KR KR1020137012534A patent/KR101453234B1/ko active IP Right Grant
- 2011-07-13 DE DE112011103794.5T patent/DE112011103794B4/de active Active
- 2011-07-13 WO PCT/JP2011/065961 patent/WO2012066819A1/ja active Application Filing
- 2011-07-13 JP JP2012544126A patent/JP5558585B2/ja active Active
- 2011-07-13 US US13/878,321 patent/US9089966B2/en active Active
- 2011-07-13 CN CN201180055387.5A patent/CN103221188B/zh active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH03270890A (ja) * | 1990-03-20 | 1991-12-03 | Canon Inc | 視覚センサを備えたワーク把持制御装置 |
JPH09277184A (ja) * | 1996-04-15 | 1997-10-28 | Ricoh Co Ltd | 画像処理方法 |
JPH11198076A (ja) * | 1998-01-05 | 1999-07-27 | Kubota Corp | 保持位置判別装置及び仕分け装置 |
JP2000263482A (ja) * | 1999-03-17 | 2000-09-26 | Denso Corp | ワークの姿勢探索方法および姿勢探索装置ならびにロボットによるワーク把持方法およびワーク把持装置 |
JP2000304509A (ja) * | 1999-04-21 | 2000-11-02 | Matsushita Electric Works Ltd | 物体特定方法及び装置 |
WO2004106009A1 (ja) * | 2003-06-02 | 2004-12-09 | Matsushita Electric Industrial Co., Ltd. | 物品操作システムおよび方法、並びに物品管理システムおよび方法 |
JP2007319997A (ja) * | 2006-06-01 | 2007-12-13 | Nissan Motor Co Ltd | 教示装置および教示方法 |
JP2010210511A (ja) * | 2009-03-11 | 2010-09-24 | Honda Motor Co Ltd | 対象物の3次元位置・姿勢認識装置及びその方法 |
Cited By (44)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014502564A (ja) * | 2010-12-30 | 2014-02-03 | ゼンロボティクス オイ | 把持位置を決定する方法、コンピュータプログラム及び装置 |
CN103659822A (zh) * | 2012-09-14 | 2014-03-26 | 株式会社安川电机 | 机器人设备 |
US9132548B2 (en) | 2013-03-18 | 2015-09-15 | Kabushiki Kaisha Yaskawa Denki | Robot picking system, control device and method of manufacturing a workpiece |
EP2781315A1 (en) * | 2013-03-18 | 2014-09-24 | Kabushiki Kaisha Yaskawa Denki | Robot picking system, control device and method of manufacturing a workpiece |
JP2014205209A (ja) * | 2013-04-11 | 2014-10-30 | 三菱電機株式会社 | ロボットシステム、及びロボットシステムの制御方法 |
JP2015009314A (ja) * | 2013-06-28 | 2015-01-19 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
US10675763B2 (en) | 2013-06-28 | 2020-06-09 | Canon Kabushiki Kaisha | Information processing apparatus for determining interference between object and grasping unit, information processing method, and storage medium |
JP2015085475A (ja) * | 2013-10-31 | 2015-05-07 | キヤノン株式会社 | 情報処理装置、情報処理方法 |
JP2015100866A (ja) * | 2013-11-22 | 2015-06-04 | 三菱電機株式会社 | ロボットシミュレーション装置、プログラム、記録媒体及び方法 |
JP2015145055A (ja) * | 2014-02-04 | 2015-08-13 | セイコーエプソン株式会社 | ロボット、ロボットシステム、制御装置および制御方法 |
US9694499B2 (en) | 2014-09-16 | 2017-07-04 | Fanuc Corporation | Article pickup apparatus for picking up randomly piled articles |
JP2016132086A (ja) * | 2015-01-22 | 2016-07-25 | 三菱電機株式会社 | ワーク供給装置およびワーク把持姿勢計算方法 |
JP2016221647A (ja) * | 2015-06-02 | 2016-12-28 | キヤノン株式会社 | 把持制御装置、把持制御方法及びプログラム |
US9620244B1 (en) | 2015-11-23 | 2017-04-11 | International Business Machines Corporation | Determining categories for memory fail conditions |
US9401222B1 (en) | 2015-11-23 | 2016-07-26 | International Business Machines Corporation | Determining categories for memory fail conditions |
US10603788B2 (en) | 2016-09-05 | 2020-03-31 | Fanuc Corporation | Robot simulation apparatus |
WO2018123087A1 (ja) * | 2016-12-28 | 2018-07-05 | オムロン株式会社 | シミュレーション装置、ロボット、シミュレーション方法、及びそのプログラム |
JP2018103339A (ja) * | 2016-12-28 | 2018-07-05 | オムロン株式会社 | シミュレーション装置、ロボット、シミュレーション方法、及びそのプログラム |
JP2018144144A (ja) * | 2017-03-03 | 2018-09-20 | 株式会社キーエンス | 画像処理装置、画像処理方法、及びコンピュータプログラム |
JP2021185016A (ja) * | 2017-03-06 | 2021-12-09 | 株式会社Fuji | 画像処理装置 |
JP7197653B2 (ja) | 2017-03-06 | 2022-12-27 | 株式会社Fuji | 画像処理装置 |
JP2019063984A (ja) * | 2017-10-02 | 2019-04-25 | キヤノン株式会社 | 情報処理装置、方法及びロボットシステム |
JP2019126886A (ja) * | 2018-01-25 | 2019-08-01 | 株式会社リコー | 情報処理システム、ピッキング箇所特定方法及びプログラム |
JP7073742B2 (ja) | 2018-01-25 | 2022-05-24 | 株式会社リコー | 情報処理システム、ピッキング箇所特定方法及びプログラム |
JP2019153605A (ja) * | 2018-02-28 | 2019-09-12 | オムロン株式会社 | 落下ばらつき予測装置、方法、プログラム、及びロボットシステム |
JP2018126862A (ja) * | 2018-05-14 | 2018-08-16 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
JP7264247B2 (ja) | 2019-07-01 | 2023-04-25 | オムロン株式会社 | 情報処理装置及び情報処理方法 |
JPWO2021001882A1 (ja) * | 2019-07-01 | 2021-01-07 | ||
WO2021001882A1 (ja) * | 2019-07-01 | 2021-01-07 | オムロン株式会社 | 情報処理装置及び情報処理方法 |
JPWO2021029064A1 (ja) * | 2019-08-15 | 2021-02-18 | ||
US12097615B2 (en) | 2019-08-15 | 2024-09-24 | Omron Corporation | Information processing device and information processing method |
JP7310895B2 (ja) | 2019-08-15 | 2023-07-19 | オムロン株式会社 | 情報処理装置及び情報処理方法 |
WO2021029064A1 (ja) * | 2019-08-15 | 2021-02-18 | オムロン株式会社 | 情報処理装置及び情報処理方法 |
JPWO2021053750A1 (ja) * | 2019-09-18 | 2021-03-25 | ||
WO2021053750A1 (ja) * | 2019-09-18 | 2021-03-25 | 株式会社Fuji | 作業ロボットおよび作業システム |
JP7163506B2 (ja) | 2019-09-18 | 2022-10-31 | 株式会社Fuji | 作業ロボットおよび作業システム |
JP2020049651A (ja) * | 2019-12-11 | 2020-04-02 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、コンピュータプログラム |
JP7066671B2 (ja) | 2019-12-11 | 2022-05-13 | キヤノン株式会社 | 干渉判定装置、干渉判定方法、プログラム及びシステム |
JP2021171885A (ja) * | 2020-04-27 | 2021-11-01 | パナソニックIpマネジメント株式会社 | 保持装置、制御方法、制御プログラム、制御システム、およびエンドエフェクタ |
JP2022087396A (ja) * | 2020-12-01 | 2022-06-13 | 株式会社エクサウィザーズ | アノテーション支援装置、方法、及びプログラム |
JP6878727B1 (ja) * | 2020-12-01 | 2021-06-02 | 株式会社エクサウィザーズ | アノテーション支援装置、方法、及びプログラム |
CN112936257A (zh) * | 2021-01-22 | 2021-06-11 | 熵智科技(深圳)有限公司 | 一种工件抓取方法、装置、计算机设备及存储介质 |
WO2023013743A1 (ja) * | 2021-08-04 | 2023-02-09 | 京セラ株式会社 | 保持パラメータ推定装置及び保持パラメータ推定方法 |
WO2024023934A1 (ja) * | 2022-07-26 | 2024-02-01 | ファナック株式会社 | ワーク取出し装置、ワーク取出し方法及び制御装置 |
Also Published As
Publication number | Publication date |
---|---|
CN103221188A (zh) | 2013-07-24 |
DE112011103794B4 (de) | 2019-01-24 |
KR20130102080A (ko) | 2013-09-16 |
JPWO2012066819A1 (ja) | 2014-05-12 |
US9089966B2 (en) | 2015-07-28 |
JP5558585B2 (ja) | 2014-07-23 |
KR101453234B1 (ko) | 2014-10-22 |
CN103221188B (zh) | 2016-08-03 |
US20130211593A1 (en) | 2013-08-15 |
DE112011103794T5 (de) | 2013-10-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5558585B2 (ja) | ワーク取り出し装置 | |
US11772267B2 (en) | Robotic system control method and controller | |
US11383380B2 (en) | Object pickup strategies for a robotic device | |
CN108399639B (zh) | 基于深度学习的快速自动抓取与摆放方法 | |
US10894324B2 (en) | Information processing apparatus, measuring apparatus, system, interference determination method, and article manufacturing method | |
JP6807949B2 (ja) | 干渉回避装置 | |
CN110640730B (zh) | 生成用于机器人场景的三维模型的方法和系统 | |
JP5788460B2 (ja) | バラ積みされた物品をロボットで取出す装置及び方法 | |
JP7481427B2 (ja) | 取り出しシステム及び方法 | |
JP2014161965A (ja) | 物品取り出し装置 | |
JP2014205209A (ja) | ロボットシステム、及びロボットシステムの制御方法 | |
US20230044001A1 (en) | Systems and methods for object detection | |
JP6945209B1 (ja) | 物体検出のために安全ボリュームリストを生成するための方法および計算システム | |
JP2020121352A (ja) | 物体把持システム | |
JP7398662B2 (ja) | ロボット多面グリッパアセンブリ及びその操作方法 | |
CN113808201B (zh) | 一种目标物体检测方法及引导抓取方法 | |
CN112338922A (zh) | 一种五轴机械臂抓取放置方法及相关装置 | |
JP7391342B2 (ja) | 計算システム、方法及び非一時的コンピュータ可読媒体 | |
US20240198543A1 (en) | Automatic gripper fingertip design to reduce leftover in random bin picking applications | |
US20230241771A1 (en) | Object placement | |
CN113219900B (zh) | 基于相机生成的图像信息执行运动规划的方法和计算系统 | |
Mohammad et al. | Developing an Object Detection and Gripping Mechanism Algorithm using Machine Learning | |
CN115256368A (zh) | 用于控制机器人的手爪抓取对象的控制装置和方法 | |
CN118617414A (zh) | 一种基于双臂机器人侧向抓取箱体方法、装置、设备和介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11841701 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2012544126 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 13878321 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 20137012534 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1120111037945 Country of ref document: DE Ref document number: 112011103794 Country of ref document: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 11841701 Country of ref document: EP Kind code of ref document: A1 |