WO2023145901A1 - 推定装置、推定方法、及び制御装置 - Google Patents
推定装置、推定方法、及び制御装置 Download PDFInfo
- Publication number
- WO2023145901A1 WO2023145901A1 PCT/JP2023/002712 JP2023002712W WO2023145901A1 WO 2023145901 A1 WO2023145901 A1 WO 2023145901A1 JP 2023002712 W JP2023002712 W JP 2023002712W WO 2023145901 A1 WO2023145901 A1 WO 2023145901A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- holding
- information
- control unit
- unit
- acquisition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Program-controlled manipulators
- B25J9/16—Program controls
- B25J9/1628—Program controls characterised by the control loop
- B25J9/163—Program controls characterised by the control loop learning, adaptive, model based, rule based expert control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Program-controlled manipulators
- B25J9/16—Program controls
- B25J9/1694—Program controls characterised by use of sensors other than normal servo-feedback from position, speed or acceleration sensors, perception control, multi-sensor controlled systems, sensor fusion
- B25J9/1697—Vision controlled systems
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—Two-dimensional [2D] image generation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—Two-dimensional [2D] image generation
- G06T11/60—Creating or editing images; Combining images with text
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- 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
- G05B2219/39484—Locate, reach and grasp, visual guided grasping
-
- 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
- G05B2219/39536—Planning of hand motion, grasping
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Definitions
- the present disclosure relates to an estimation device, an estimation method, and a control device.
- Non-Patent Document 1 a neural network that uses an image as input to infer the gripping position.
- An estimation device includes an acquisition unit and a control unit.
- the acquisition unit acquires information about a holding object held by the holding unit.
- the control unit estimates a holding position at which the holding unit holds the holding object, considering an acquisition point of the information on the holding object.
- An estimation device includes an acquisition unit and a control unit.
- the acquisition unit acquires information about a holding object held by the holding unit.
- the control section estimates a holding position at which the holding section holds the holding target based on information about the holding target.
- the control unit calculates a difference between a direction from an acquisition point of information on the holding object to the holding object and a direction in which the holding unit holds the holding object as a direction deviation.
- the control unit estimates a holding position at which the holding unit holds the holding object based on information about the holding object when the direction deviation satisfies a predetermined condition.
- the control unit estimates a holding position at which the holding unit holds the holding object based on the acquired point and information about the holding object when the direction deviation does not satisfy the predetermined condition.
- An estimation method includes acquiring information about a holding target held by a holding unit.
- the estimation method includes estimating a holding position at which the holding part holds the holding object, based on the information on the holding object, considering an acquisition point of the information on the holding object.
- An estimation method includes acquiring information about a holding target held by a holding unit.
- the estimation method includes calculating a difference between a direction from an acquisition point of information on the holding object to the holding object and a direction in which the holding unit holds the holding object as a direction deviation.
- the estimating method includes estimating a holding position where the holding part holds the holding object based on information about the holding object when the direction deviation satisfies a predetermined condition.
- the estimation method includes estimating a holding position at which the holding unit holds the holding object based on the acquisition point and information about the holding object when the direction deviation does not satisfy the predetermined condition. include.
- a control device causes a holding unit to hold a holding object at a holding position estimated by the estimation device or at a holding position estimated by executing the estimation method.
- FIG. 1 is a schematic diagram showing a configuration example of a robot control system according to an embodiment
- FIG. FIG. 4 is a cross-sectional view showing an example of a holding position of an object held by a hand
- 1 is a block diagram showing a configuration example of a robot control system according to an embodiment
- FIG. FIG. 4 is a diagram showing an example of an inference model of a holding position
- FIG. 10 is a diagram showing an example of estimation results of a holding position when the object to be held is viewed from directly above (when direction deviation is small);
- 10 is a diagram showing an example of a holding position estimation result when an object to be held is photographed obliquely (when the direction deviation is large); 4 is a flow chart showing an example procedure of an estimation method according to an embodiment; 4 is a flow chart showing an example procedure of an estimation method including direction deviation determination;
- a robot control system 1 includes a robot 2, an information acquisition unit 4, a control device 10, and an estimation device 20.
- a control device 10 controls the robot 2 .
- the estimating device 20 estimates a holding position 82 of the holding object 8 by the robot 2 and outputs it to the control device 10 .
- the robot 2 holds the object 8 to be held on the work start table 6 . That is, the control device 10 controls the robot 2 to hold the holding object 8 on the work start table 6 .
- the robot 2 may move the object 8 to be held from the work start table 6 to the work target table 7 .
- the robot 2 operates inside the operating range 5 .
- the robot 2 includes an arm 2A and a holding section 2B.
- the arm 2A may be configured as, for example, a 6-axis or 7-axis vertical articulated robot.
- the arm 2A may be configured as a 3-axis or 4-axis horizontal articulated robot or SCARA robot.
- the arm 2A may be configured as a 2-axis or 3-axis Cartesian robot.
- Arm 2A may be configured as a parallel link robot or the like.
- the number of shafts forming the arm 2A is not limited to the illustrated one.
- the robot 2 has an arm 2A connected by a plurality of joints and operates by driving the joints.
- the holding part 2B may be, for example, a hand or a suction part. In this embodiment, it is assumed that the holding part 2B is the hand 2B.
- the holding part 2B is not limited to a hand, and may be, for example, a suction part having a suction nozzle for sucking the object 8 to be held.
- holding the holding object 8 with the hand 2B may be referred to as gripping.
- the held object 8 may be referred to as a gripped object 8 .
- "holding” used in the following description may be interpreted as “holding”.
- “gripping position 82" used in the description below may be interpreted as "holding position 82".
- the hand 2B may include, for example, a gripper configured to grip the gripping target 8.
- the gripper may have at least one finger 2C.
- the gripper fingers 2C may have one or more joints.
- the fingers 2C of the gripper may have a suction portion that grasps the grasped object 8 by suction.
- the hand 2B may be configured as one finger 2C having a suction portion.
- the hand 2B may be configured as two or more fingers 2C that pinch and grip the gripping target 8 .
- the hand 2B is not limited to these examples, and may be configured to perform various other operations.
- the hand 2B includes a gripper with two fingers 2C.
- the hand 2B grips the gripping object 8 with two fingers 2C.
- a position at which the gripping object 8 is gripped is represented as a gripping position 82 .
- the gripping position 82 is the midpoint (horizontal and vertical direction).
- the control device 10 can control the position of the hand 2B by operating the arm 2A of the robot 2 .
- the hand 2B may have an axis that serves as a reference for the direction in which the hand 2B acts on the grasped object 8. As shown in FIG. When the hand 2B has an axis, the control device 10 can control the direction of the axis of the hand 2B by operating the arm 2A.
- the control device 10 controls the start and end of the action of the hand 2B acting on the gripped object 8 .
- the control device 10 can move or process the grasped object 8 by controlling the position of the hand 2B or the direction of the axis of the hand 2B and controlling the operation of the hand 2B. In the configuration illustrated in FIG.
- the control device 10 controls the robot 2 so that the hand 2B grips the gripping object 8 on the work start table 6 and moves the hand 2B to the work target table 7 .
- the control device 10 controls the robot 2 so that the hand 2B releases the gripped object 8 on the work target table 7 . By doing so, the control device 10 can move the grasped object 8 from the work start table 6 to the work target table 7 by the robot 2 .
- the control device 10 may be configured including at least one processor.
- the processor may execute programs that implement various functions of the controller 10 .
- a processor may be implemented as a single integrated circuit.
- An integrated circuit is also called an IC (Integrated Circuit).
- a processor may be implemented as a plurality of communicatively coupled integrated and discrete circuits. Processors may be implemented based on various other known technologies.
- the control device 10 may include a storage unit.
- the storage unit may include an electromagnetic storage medium such as a magnetic disk, or may include a memory such as a semiconductor memory or a magnetic memory.
- the storage unit stores various information.
- the storage unit stores programs and the like executed by the control device 10 .
- the storage unit may be configured as a non-transitory readable medium.
- the storage unit may function as a work memory for the control device 10 . At least part of the storage unit may be configured separately from the control device 10 .
- the estimation device 20 includes a control section 22 , an acquisition section 24 and a display section 26 .
- the acquisition unit 24 acquires information about the grasped object 8 .
- the control unit 22 estimates the position at which the hand 2 ⁇ /b>B is to grip the gripping object 8 based on the information about the gripping object 8 , and outputs the estimated position to the control device 10 .
- the display unit 26 may display the estimation result of the gripping position 82 of the gripping object 8 .
- the acquisition unit 24 may receive an input for correcting the gripping position 82 from the user who has viewed the estimation result of the gripping position 82 .
- the control unit 22 may correct the gripping position 82 based on the user's input and output the corrected gripping position 82 to the control device 10 .
- the control unit 22 may estimate the gripping position 82 using an inference model.
- the inference model can be configured as a trained model 30, as illustrated in FIG.
- the trained model 30 can be expressed as a model connecting the first model 31 and the second model 32 .
- the first model 31 is configured to output the result of extracting the feature amount of the input information.
- the feature quantity represents features of the appearance such as edges or patterns of the gripping object 8, for example.
- the first model 31 may be configured including a CNN (Convolution Neural Network) having multiple layers.
- the first model 31 may include, for example, convolution and pooling.
- the second model 32 is configured to make a predetermined judgment on the input information based on the output of the first model 31 . Specifically, the second model 32 may output an estimation result of the gripping position 82 of the gripping object 8 included in the input information based on the feature amount output by the first model 31 .
- the second model 32 may include a fully connected layer that processes the result of feature extraction by the first model 31 .
- the inference model may include a convolutional layer that receives input of information about the grasped object 8 and a fully connected layer that processes the output of the convolutional layer and outputs the inference result of the grasped position 82 .
- Fully connected layers may include layers that consider acquisition point information.
- the control unit 22 may include at least one processor to provide control and processing power to perform various functions.
- the processor may execute programs that implement various functions of the controller 22 .
- a processor may be implemented as a single integrated circuit.
- An integrated circuit is also called an IC (Integrated Circuit).
- a processor may be implemented as a plurality of communicatively coupled integrated and discrete circuits. Processors may be implemented based on various other known technologies.
- the control unit 22 may include a storage unit.
- the storage unit may include an electromagnetic storage medium such as a magnetic disk, or may include a memory such as a semiconductor memory or a magnetic memory.
- the storage unit stores various information.
- the storage unit stores programs and the like executed by the control unit 22 .
- the storage unit may be configured as a non-transitory readable medium.
- the storage section may function as a work memory for the control section 22 . At least part of the storage section may be configured separately from the control section 22 .
- the acquisition unit 24 may be configured including a communication device configured to be capable of wired or wireless communication.
- a communication device may be configured to be able to communicate with communication schemes based on various communication standards.
- a communication device may be configured according to known communication technologies.
- the acquisition unit 24 may include an input device that receives input of information, data, etc. from the user.
- the input device may include, for example, a touch panel or touch sensor, or a pointing device such as a mouse.
- the input device may be configured including physical keys.
- the input device may include an audio input device such as a microphone.
- the display unit 26 includes a display device that displays information or data to the user.
- the display device is configured to output visual information such as images or characters or graphics, for example.
- the display device may include, for example, an LCD (Liquid Crystal Display), an organic EL (Electro-Luminescence) display or an inorganic EL display, or a PDP (Plasma Display Panel).
- the display device is not limited to these displays, and may be configured to include other various types of displays.
- the display device may include a light emitting device such as an LED (Light Emission Diode) or an LD (Laser Diode).
- the display device may be configured including other various devices.
- the information acquisition unit 4 acquires information on the grasped object 8 .
- the information acquisition unit 4 may be configured including a camera.
- a camera as the information acquisition unit 4 captures an image of the grasped object 8 as information of the grasped object 8 .
- the information acquisition unit 4 may be configured including a depth sensor.
- a depth sensor as the information acquisition unit 4 acquires depth data of the grasped object 8 .
- the depth data may be converted into point cloud information of the gripped object 8 .
- a control device 10 controls motions of the robot 2 .
- the estimation device 20 estimates the position at which the hand 2B of the robot 2 grips the gripping target 8 when the hand 2B of the robot 2 operates to grip the gripping target 8 .
- the control device 10 determines a position for the hand 2B to grip the gripping target 8 based on the estimation result by the estimating device 20, and controls the arm of the robot 2 so that the hand 2B grips the gripping target 8 at the determined position. Controls hand 2A or hand 2B.
- An operation example of the estimating device 20 will be described below.
- the acquiring unit 24 of the estimating device 20 acquires information about the grasped object 8 grasped by the hand 2B from the information obtaining unit 4.
- Information about the grasped object 8 is also referred to as object information.
- the information about the grasped object 8 includes an image obtained by capturing the grasped object 8 from the information acquisition unit 4, distance data from the information acquisition unit 4 to the grasped object 8, or the like.
- the acquisition unit 24 outputs information (target information) regarding the gripped object 8 acquired from the information acquisition unit 4 to the control unit 22 of the estimation device 20 .
- the control unit 22 acquires information (object information) on the gripping object 8 by the acquisition unit 24 .
- control unit 22 determines the position ( Estimate the gripping position 82).
- the control unit 22 may input information (target information) about the grasped object 8 to an inference model such as the learned model 30 and estimate an inference result output from the inference model as the grasped position 82 .
- the grip position 82 may be represented by the coordinates of each of the two fingers 2C.
- the gripping position 82 may be represented by the coordinates of the middle point of the two fingers 2C and the rotation angle of the hand 2B. Even if the hand 2B has three or more fingers 2C, the gripping position 82 may be represented by the coordinates of each finger 2C, or the average coordinates of the coordinates of each finger 2C and the rotation angle of the hand 2B. may be represented by If the information about the grasped object 8 is two-dimensionally mapped information such as an image or distance data, the grasped position 82 may be expressed as coordinates within the plane on which the information is mapped. If the information about the gripping object 8 is three-dimensional information, the gripping position 82 may be represented as three-dimensional coordinates.
- the position at which information (object information) on the grasped object 8 is acquired is also referred to as an acquisition point.
- the information about acquired points is also referred to as acquired point information.
- Acquisition point information may include information about the direction in which the acquisition point is located when viewed from the gripped object 8 .
- the direction in which the acquisition point is located when viewed from the grasped object 8 is also referred to as the acquisition direction.
- Acquisition point information may include information about the direction in which the hand 2B grips the gripping target 8 (the direction or posture of the hand 2B when gripping the gripping target 8).
- the direction in which the hand 2B grips the gripping object 8 is also referred to as the gripping direction.
- the gripping direction may be set in advance, for example, in the direction of weight or in the direction perpendicular to the work surface of the work table such as the work start table 6 or the work target table 7 .
- Each of the acquisition direction and the gripping direction may be represented by two types of angles in a polar coordinate system, or may be represented as a unit vector in a three-dimensional space.
- Acquisition point information may include information representing the relative relationship between the acquisition direction and the gripping direction.
- the relative relationship between the acquisition direction and the grasping direction can be specifically expressed as the difference between the angles representing the acquiring direction and the grasping direction, respectively.
- the control unit 22 may acquire acquisition point information from the information acquisition unit 4 .
- Acquisition point information may be information specifying a default acquisition point.
- Acquisition point information may be information expressing the direction (acquisition direction) in which information about the gripped object 8 is acquired in terms of roll, pitch, and yaw with the position of the gripped object 8 as the origin, or information represented by a quaternion.
- a quaternion is a format that represents a posture rotated by a predetermined angle about a direction vector as a rotation axis.
- the control unit 22 may acquire acquisition point information by generating acquisition point information based on target information. For example, when an image is acquired as target information, acquisition point information may be generated using a P3P algorithm, which is a type of Perspective-n-Point algorithm, using markers whose positions on three-dimensional coordinates are known. .
- the acquisition point information is obtained by projecting the reference marker together with the grasped object 8, and the shape of the reference marker on the image and the reference marker when the reference marker and the camera, which is the information acquisition unit 4, face each other.
- the acquisition point information is obtained by determining the reference posture of the grasped object 8, and the outline of the grasped object 8 appearing in the image as the target information, or the distance data from the information acquisition unit 4 to the grasped object 8, or the like. It may be generated by comparing with the information in the reference posture.
- the difference between the acquisition direction and the gripping direction is also called direction deviation.
- the control unit 22 may calculate the difference between the acquisition direction and the gripping direction as a value representing the direction deviation.
- the control unit 22 may calculate the angle difference as a value representing the direction deviation.
- the control unit 22 may calculate the angle formed by the unit vectors of the acquisition direction and the gripping direction as a value representing the direction deviation. An inner product or outer product of each unit vector may be calculated.
- the control unit 22 may estimate the gripping position 82 at which the hand 2B grips the gripping target 8 based on the information (target information) on the gripping target 8 .
- the control unit 22 estimates a gripping position 82 at which the hand 2B grips the gripping object 8 based on the acquired point and information (object information) about the gripping object 8. good.
- the predetermined condition may include that the value representing the direction deviation (the difference between the acquisition direction and the gripping direction) is within a predetermined range. Specifically, the predetermined condition may include that the angular difference calculated as the misorientation is less than an angular threshold.
- the angle threshold may be set at 10 degrees, for example.
- the predetermined condition may include that the absolute value of the cross product of the unit vectors calculated as the direction deviation is less than the cross product threshold.
- the predetermined condition may include that the inner product of the unit vectors calculated as the direction deviation is greater than the inner product threshold.
- the control unit 22 can estimate the grip position 82 using the learned model 30 as an inference model.
- the second model 32 may be configured to output an estimation result of the gripping position 82 considering the acquisition point information.
- the second model 32 may be configured to receive input of acquisition point information.
- the second model 32 may be configured to process the feature quantity output from the first model 31 in three steps.
- the second model 32 may have a layer that performs preprocessing as a first stage of processing.
- the second model 32 may have layers that perform a combination process as a second stage process.
- the second model 32 may have a layer that performs output processing as a third stage of processing.
- the second model 32 may be configured to receive input of acquired point information in the layer that executes the combining process as the process of the second stage. That is, as the second-stage process of the second model 32, a process of combining the feature amount of the grasped object 8 and the acquisition point information may be executed. In this case, the feature amount of the grasped object 8 output from the first model 31 is processed so as to combine acquisition point information in the second stage of combining processing.
- the processing of the second model 32 is divided into three stages, but it may be divided into four or more stages, two stages, or a plurality of stages. It doesn't have to be.
- the gripping position 82 may be represented by the coordinates of the middle point of the two fingers 2C in the image of the target information. Furthermore, the gripping position 82 may be represented by information specifying that the two fingers 2C of the hand 2B are aligned along the Y-axis direction.
- the acquisition direction is tilted toward the positive direction of the Y-axis with respect to the Z-axis.
- an image is acquired as target information.
- the grasped object 8 appears in the image in a tilted form.
- an estimated gripping position 50F represented by a two-dot chain line is estimated as the gripping position 82.
- the hand 2B grips the gripping target 8 from the actual gripping direction, the hand 2B may be unable or difficult to grip the gripping target 8 at the estimated gripping position 50F.
- the estimated grasped position 50T is estimated as the grasped position 82. be.
- the estimated gripping position 50T appears to be displaced from the gripped object 8 appearing in the image.
- the hand 2B grips the gripping target 8 from the actual gripping direction
- the hand 2B tends to grip the gripping target 8 at the estimated gripping position 50T.
- the control unit 22 determines that the gripping of the gripping target 8 by the hand 2B is more successful than in the case of estimating the gripping position 82 without considering the acquisition points. rate can be increased.
- the calculation load of the inference model when outputting the estimation result of the gripping position 82 considering the acquisition point is larger than the calculation load of the inference model when outputting the estimation result of the gripping position 82 without considering the acquisition point.
- the gripping position 82 is estimated with high accuracy without considering the acquired points as illustrated in FIG. 5, the acquired points may not be considered.
- the acquired points may be taken into consideration when the estimation accuracy of the gripped position 82 is degraded if the acquired points are not taken into account.
- the control unit 22 of the estimating device 20 may execute an estimating method including the procedures of the flowchart illustrated in FIG.
- the estimation method may be implemented as an estimation program that is executed by a processor that configures the control unit 22 .
- the estimation program may be stored on a non-transitory computer-readable medium.
- the control unit 22 acquires target information from the information acquisition unit 4 (step S1).
- the control unit 22 may acquire an image of the grasped object 8 as the object information.
- the control unit 22 acquires acquisition point information (step S2).
- the control unit 22 may acquire acquisition point information from the information acquisition unit 4 .
- the control unit 22 may acquire acquisition point information by generating acquisition point information based on target information.
- the control unit 22 estimates the gripping position 82 based on the target information (step S3).
- the control unit 22 inputs the target information to the inference model, and estimates the gripping position 82 by acquiring the estimation result of the gripping position 82 output from the inference model.
- the control unit 22 inputs the acquisition point information to the inference model, thereby causing the inference model to output the estimation result of the gripping position 82 considering the acquisition point information. you can
- the control unit 22 may cause the inference model to output the estimation result of the gripping position 82 considering the acquisition point information by setting a parameter that instructs the inference model to make an estimation considering the acquisition point information. .
- the control unit 22 causes the display unit 26 to display the estimation result of the gripping position 82 (step S4).
- the control unit 22 determines whether correction information for the estimation result of the gripping position 82 has been acquired (step S5). Specifically, the control unit 22 may determine that the correction information is acquired when the correction information is input to the acquisition unit 24 by the user who has seen the estimation result of the gripping position 82 . If the control unit 22 does not acquire the correction information (step S5: NO), the process proceeds to step S7. When the correction information is acquired (step S5: YES), the control unit 22 corrects the gripping position 82 based on the correction information (step S6).
- the control unit 22 controls the gripping operation of the robot 2 so that the hand 2B grips the gripping object 8 at the estimated gripping position 82 or the corrected gripping position 82 (step S7). After executing the procedure of step S7, the control unit 22 ends the execution of the flowchart of FIG.
- the control unit 22 may execute an estimation method including the procedure of the flowchart illustrated in FIG.
- the control unit 22 acquires target information from the information acquisition unit 4 (step S11).
- the control unit 22 acquires acquisition point information (step S12).
- the procedures of steps S11 and S12 may be performed as the same or similar procedures as the procedures of steps S1 and S2 of FIG.
- the control unit 22 calculates the direction deviation (step S13). Specifically, the control unit 22 calculates the difference between the direction from the acquired point to the grasped object 8 and the direction in which the hand 2B grasps the grasped object 8 as the direction deviation. The control unit 22 determines whether the direction deviation satisfies a predetermined condition (step S14).
- step S14 When the direction deviation satisfies a predetermined condition (step S14: YES), the control unit 22 estimates the gripping position 82 without considering the acquired points (step S15). Specifically, the control unit 22 inputs the target information to the inference model but does not input the acquisition point information, and acquires the estimation result of the gripping position 82 output from the inference model. After executing the procedure of step S15, the control unit 22 proceeds to the procedure of step S4 in FIG.
- step S16 the control unit 22 estimates the gripping position 82 considering the acquired points. Specifically, the control unit 22 inputs acquisition point information together with target information to the inference model, and acquires an estimation result of the gripping position 82 output from the inference model. After executing the procedure of step S16, the control unit 22 proceeds to the procedure of step S4 in FIG.
- the accuracy of estimating the gripping position 82 can be improved even when the gripping direction and the acquisition direction do not match. As a result, gripping stability can be enhanced.
- the control device 10 or the estimation device 20 may be configured as a server device.
- the server device may be configured including at least one computer.
- the server device may be configured to allow multiple computers to execute parallel processing.
- the server device does not need to be configured including a physical enclosure, and may be configured based on virtualization technology such as a virtual machine or a container orchestration system.
- the server device may be configured using a cloud service. When the server device is configured using cloud services, it can be configured by combining managed services. That is, the functions of the control device 10 can be implemented as cloud services.
- the server device may comprise at least one server group.
- the server group functions as the control unit 22 .
- the number of server groups may be one or two or more. When the number of server groups is one, functions realized by one server group include functions realized by each server group.
- Each server group is communicably connected to each other by wire or wirelessly.
- control device 10 or the estimating device 20 is described as one configuration in each of FIGS. 1 and 2, multiple configurations can be regarded as one system and operated as necessary. That is, the control device 10 or the estimation device 20 is configured as a platform with variable capacity. By using a plurality of configurations as the control device 10 or the estimating device 20, even if one configuration becomes inoperable in the event of an unforeseen event such as a natural disaster, the system continues to operate using the other configurations. . In this case, each of the plurality of components is connected by a line, whether wired or wireless, and configured to be able to communicate with each other. These multiple configurations may be built across cloud services and on-premises environments.
- control device 10 or the estimating device 20 is connected to the robot 2 by, for example, a wired or wireless communication line.
- the control device 10, the estimating device 20, or the robot 2 are equipped with communication devices that use standard protocols with each other, and are capable of two-way communication.
- the control device 10 may control the robot 2 so that the hand 2B grips the gripping target 8 at the gripping position 82 estimated by the estimating device 20 .
- the control device 10 may control the robot 2 so that the hand 2B grips the gripping target 8 at the gripping position 82 estimated by executing the estimation method.
- the control device 10 and the estimation device 20 may be configured integrally.
- control unit 22 inputs target information into the inference model to estimate the gripping position 82 .
- the control unit 22 may further calculate the grip position 82 by processing the target information with a rule-based algorithm.
- Rule-based algorithms may include, for example, template matching and may include processing with maps.
- the control unit 22 may generate, for example, a map that specifies the form of the hand 2B (interval between the fingers 2C, thickness or width of the fingers 2C, etc.).
- the control unit 22 may generate a rule map representing suitability of each part of the gripping object 8 as the gripping position 82 .
- the rule map is, for example, a surrounding environment map representing the surrounding environment of the gripped object 8, an object map representing the characteristics (shape, center of gravity, material, etc.) of the gripped object 8, or the surfaces of the hand 2B and the gripped object 8. may include a contact map or the like representing rules based on relationships with the state of the .
- the control unit 22 may compare the gripping position 82 calculated by the rule-based algorithm and the gripping position 82 estimated by the inference model. Based on the comparison result, the control unit 22 may check whether the estimation accuracy of the gripping position 82 estimation result exceeds a predetermined accuracy. Further, the control unit 22 may correct the estimation result of the gripping position 82 based on the comparison result, and determine the corrected estimation result as the gripping position 82 . In other words, the control unit 22 inspects or corrects the estimation result of the gripping position 82 based on the inference model based on the result obtained by processing the information (target information) about the gripping target 8 with a rule-based algorithm. good.
- the accuracy of the gripping position 82 can be enhanced by using the results obtained by the rule-based algorithm. As a result, gripping stability can be enhanced.
- the control unit 22 may cause the display unit 26 to display the estimation result of the gripping position 82 of the gripping object 8 .
- the control unit 22 may cause the display unit 26 to display a superimposed image in which an image representing the estimation result of the gripping position 82 of the gripping object 8 is superimposed on an image representing information about the gripping object 8 .
- the user can determine how to correct the gripping position 82 by viewing the superimposed image.
- control unit 22 may convert an image of the grasped object 8 captured from the acquisition point into an image assuming that the grasped object 8 is captured from the grasping direction, and cause the display unit 26 to display the image.
- the control unit 22 may convert the image representing the information about the grasped object 8 into an image obtained when the information about the grasped object 8 is acquired from the direction in which the hand 2B grasps the grasped object 8 .
- the control unit 22 may transform the image by perspective projection transformation.
- the control unit 22 may transform the image by coordinate transformation of point cloud information obtained by photographing the grasped object 8 using an RGB-D camera. RGB-D cameras are configured to acquire both color images (or grayscale images) and depth data. By converting and displaying the image, it becomes easier for the user to determine how the grip position 82 should be corrected.
- the control unit 22 may receive an input for the user to correct the gripping position 82 based on the superimposed image by the acquisition unit 24 .
- the control unit 22 may correct the gripping position 82 based on the user's input. By doing so, the accuracy of the grip position 82 can be enhanced. As a result, gripping stability can be enhanced.
- the control unit 22 may re-learn the inference model based on the input for correcting the gripping position 82 .
- the control unit 22 accumulates inputs for correcting the gripping position 82, and when the number of accumulated correction inputs reaches a predetermined number or more, the accumulated correction inputs may be collectively applied to the re-learning of the inference model.
- the control unit 22 may correct the gripped position 82 by processing the estimation result of the gripped position 82 output from the inference model with a correction filter.
- the control unit 22 may update the correction filter based on the input for correcting the gripping position 82 .
- the control unit 22 may accumulate inputs for correcting the grip position 82, and apply the accumulated correction inputs collectively to update the correction filter when the number of accumulated correction inputs reaches a predetermined number or more.
- a storage medium on which the program is recorded for example, an optical disk, a magneto-optical disk , CD-ROM, CD-R, CD-RW, magnetic tape, hard disk, memory card, etc.
- the implementation form of the program is not limited to an application program such as an object code compiled by a compiler or a program code executed by an interpreter. good.
- the program may or may not be configured so that all processing is performed only in the CPU on the control board.
- the program may be configured to be partially or wholly executed by another processing unit mounted on an expansion board or expansion unit added to the board as required.
- Embodiments according to the present disclosure are not limited to any specific configuration of the embodiments described above. Embodiments of the present disclosure extend to all novel features or combinations thereof described in the present disclosure or to all novel method or process steps or combinations thereof described. be able to.
- robot control system (4: information acquisition unit, 5: robot operating range, 6: work start table, 7: work target table, 10: control device) 2 robot (2A: arm, 2B: hand, 2C: finger) 8 gripping object (82: gripping position) 20 estimation device (22: control unit, 24: acquisition unit, 26: display unit) 30 trained models (31: first model, 32: second model) 50F, 50T Estimated grip position
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Manipulator (AREA)
Abstract
Description
図1、図2及び図3に示されるように、本開示の一実施形態に係るロボット制御システム1は、ロボット2と、情報取得部4と、制御装置10と、推定装置20とを備える。制御装置10は、ロボット2を制御する。推定装置20は、ロボット2による保持対象物8の保持位置82を推定し、制御装置10に出力する。
ロボット2は、アーム2Aと、保持部2Bとを備える。アーム2Aは、例えば、6軸又は7軸の垂直多関節ロボットとして構成されてよい。アーム2Aは、3軸又は4軸の水平多関節ロボット又はスカラロボットとして構成されてもよい。アーム2Aは、2軸又は3軸の直交ロボットとして構成されてもよい。アーム2Aは、パラレルリンクロボット等として構成されてもよい。アーム2Aを構成する軸の数は、例示したものに限られない。言い換えれば、ロボット2は、複数の関節で接続されるアーム2Aを有し、関節の駆動によって動作する。
制御装置10は、ロボット2のアーム2Aを動作させることによって、ハンド2Bの位置を制御できる。ハンド2Bは、把持対象物8に対して作用する方向の基準となる軸を有してもよい。ハンド2Bが軸を有する場合、制御装置10は、アーム2Aを動作させることによって、ハンド2Bの軸の方向を制御できる。制御装置10は、ハンド2Bが把持対象物8に作用する動作の開始及び終了を制御する。制御装置10は、ハンド2Bの位置、又は、ハンド2Bの軸の方向を制御しつつ、ハンド2Bの動作を制御することによって、把持対象物8を動かしたり加工したりすることができる。図1に例示される構成において、制御装置10は、作業開始台6でハンド2Bに把持対象物8を把持させ、ハンド2Bを作業目標台7へ移動させるようにロボット2を制御する。制御装置10は、作業目標台7でハンド2Bに把持対象物8を解放させるようにロボット2を制御する。このようにすることで、制御装置10は、ロボット2によって把持対象物8を作業開始台6から作業目標台7へ移動させることができる。
図3に示されるように、推定装置20は、制御部22と、取得部24と、表示部26とを備える。取得部24は、把持対象物8に関する情報を取得する。制御部22は、把持対象物8に関する情報に基づいて、ハンド2Bに把持対象物8を把持させる位置を推定し、制御装置10に出力する。
情報取得部4は、把持対象物8の情報を取得する。情報取得部4は、カメラを含んで構成されてよい。情報取得部4としてのカメラは、把持対象物8の情報として把持対象物8の画像を撮影する。情報取得部4は、デプスセンサを含んで構成されてよい。情報取得部4としてのデプスセンサは、把持対象物8のデプスデータを取得する。デプスデータは、把持対象物8の点群情報に変換されてよい。
ロボット制御システム1において、制御装置10は、ロボット2の動作を制御する。推定装置20は、ロボット2のハンド2Bが把持対象物8を把持するように動作する場合に、ハンド2Bに把持対象物8を把持させる位置を推定する。制御装置10は、推定装置20による推定結果に基づいて、ハンド2Bに把持対象物8を把持させる位置を決定し、決定した位置でハンド2Bが把持対象物8を把持するようにロボット2のアーム2A又はハンド2Bを制御する。以下、推定装置20の動作例が説明される。
上述したように、制御部22は、推論モデルとして学習済みモデル30を用いて把持位置82を推定できる。第2モデル32は、取得点情報を考慮して把持位置82の推定結果を出力するように構成されてよい。具体的に、第2モデル32は、取得点情報の入力を受け付けるように構成されてよい。第2モデル32は、第1モデル31から出力された特徴量を3段階に分けて処理するように構成されてよい。例えば、第2モデル32は、第1段階の処理として前処理を実行する層を有してよい。第2モデル32は、第2段階の処理として結合処理を実行する層を有してよい。第2モデル32は、第3段階の処理として出力処理を実行する層を有してよい。
例えば図5に示されるように、取得方向と把持方向とがZ軸に沿う方向で一致するように(方向ずれが小さくなるように)、対象情報としての画像が取得されるとする。把持対象物8を2本の指2Cで把持する場合、把持位置82は、対象情報の画像において2本の指2Cの中点の座標によって表されてよい。さらに、把持位置82は、ハンド2Bの2本の指2CがY軸方向に沿って並ぶことを特定する情報によって表されてよい。
推定装置20の制御部22は、図7に例示されるフローチャートの手順を含む推定方法を実行してもよい。推定方法は、制御部22を構成するプロセッサに実行させる推定プログラムとして実現されてもよい。推定プログラムは、非一時的なコンピュータ読み取り可能な媒体に格納されてよい。
以上述べてきたように、本実施形態に係る推定装置20及び推定方法によれば、把持方向と取得方向とが一致しない場合でも、把持位置82の推定精度が高められ得る。その結果、把持安定性が高められ得る。
以下、他の実施形態が説明される。
制御装置10又は推定装置20は、サーバ装置として構成されてもよい。サーバ装置は、少なくとも1台のコンピュータを含んで構成されてよい。サーバ装置は、複数のコンピュータに並列処理を実行させるように構成されてよい。サーバ装置は、物理的な筐体を含んで構成される必要はなく、ヴァーチャルマシン又はコンテナオーケストレーションシステムなどの仮想化技術に基づいて構成されてもよい。サーバ装置は、クラウドサービスを用いて構成されてもよい。サーバ装置がクラウドサービスを用いて構成される場合、マネージドサービスを組み合わせることで構成され得る。つまり、制御装置10の機能は、クラウドサービスとして実現され得る。
上述してきたように、制御部22は、対象情報を推論モデルに入力して把持位置82を推定する。制御部22は、さらに、対象情報をルールベースのアルゴリズムで処理することによって把持位置82を算出してもよい。ルールベースのアルゴリズムは、例えばテンプレートマッチングを含んでよいし、マップを用いた処理を含んでよい。
上述したように、制御部22は、把持対象物8の把持位置82の推定結果を表示部26に表示させてよい。制御部22は、把持対象物8の把持位置82の推定結果を表す画像を把持対象物8に関する情報を表す画像に重畳した重畳画像を、表示部26に表示させてよい。ユーザは、重畳画像を視認することによって、把持位置82をどのように補正すべきか判断できる。
2 ロボット(2A:アーム、2B:ハンド、2C:指)
8 把持対象物(82:把持位置)
20 推定装置(22:制御部、24:取得部、26:表示部)
30 学習済みモデル(31:第1モデル、32:第2モデル)
50F、50T 推定把持位置
Claims (15)
- 保持部によって保持される保持対象物に関する情報を取得する取得部と、
前記保持対象物に関する情報に基づいて、前記保持対象物に関する情報の取得点を考慮し、前記保持対象物を前記保持部に保持させる保持位置を推定する制御部と
を備える推定装置。 - 保持部によって保持される保持対象物に関する情報を取得する取得部と、
前記保持対象物に関する情報に基づいて、前記保持対象物を前記保持部に保持させる保持位置を推定する制御部と
を備え、
前記制御部は、
前記保持対象物に関する情報の取得点から前記保持対象物への方向と、前記保持部が前記保持対象物を保持する方向との差を方向ずれとして算出し、
前記方向ずれが所定条件を満たす場合に、前記保持対象物に関する情報に基づいて前記保持対象物を保持部に保持させる保持位置を推定し、
前記方向ずれが前記所定条件を満たさない場合に、前記取得点と前記保持対象物に関する情報とに基づいて前記保持対象物を前記保持部に保持させる保持位置を推定する、
推定装置。 - 前記制御部は、
前記保持対象物に関する情報を推論モデルに入力し、前記推論モデルから出力される推論結果を前記保持位置として推定する、請求項1又は2に記載の推定装置。 - 前記制御部は、
前記保持対象物に関する情報をルールベースのアルゴリズムで処理して得られた結果に基づいて、前記推論モデルによる前記保持位置の推定結果を検査又は補正する、請求項3に記載の推定装置。 - 前記推論モデルは、前記保持対象物に関する情報の入力を受け付ける畳み込み層と、前記畳み込み層の出力を処理して前記保持位置の推定結果を出力する全結合層とを含み、
前記全結合層は、前記取得点を考慮する層を含む、請求項3又は4に記載の推定装置。 - 前記制御部は、
前記保持位置の推定結果を表す画像を前記保持対象物に関する情報を表す画像に重畳した重畳画像を表示する、請求項1から5までのいずれか一項に記載の推定装置。 - 前記制御部は、
前記保持対象物に関する情報を表す画像を、前記保持部が前記保持対象物を保持する方向から前記保持対象物に関する情報を取得した場合の画像に変換する、請求項6に記載の推定装置。 - 前記制御部は、
前記重畳画像に基づいてユーザが前記保持位置を補正する入力を受け付ける、請求項6又は7に記載の推定装置。 - 前記制御部は、
前記取得点を特定する取得点情報を取得する、請求項1から8までのいずれか一項に記載の推定装置。 - 前記取得点情報は、既定の取得点を特定する情報である、請求項9に記載の推定装置。
- 前記取得点情報は、前記保持対象物に関する情報を取得した方向を、前記保持対象物の位置を原点として、ロール、ピッチ及びヨーで表す情報、又は、クオータニオンで表す情報である、請求項9に記載の推定装置。
- 前記制御部は、
前記保持対象物に関する情報に基づいて前記取得点を特定する取得点情報を生成する、請求項1から8までのいずれか一項に記載の推定装置。 - 保持部によって保持される保持対象物に関する情報を取得することと、
前記保持対象物に関する情報に基づいて、前記保持対象物に関する情報の取得点を考慮し、前記保持対象物を前記保持部に保持させる保持位置を推定することと
を含む推定方法。 - 保持部によって保持される保持対象物に関する情報を取得することと、
前記保持対象物に関する情報の取得点から前記保持対象物への方向と、前記保持部が前記保持対象物を保持する方向との差を方向ずれとして算出することと、
前記方向ずれが所定条件を満たす場合に、前記保持対象物に関する情報に基づいて前記保持対象物を保持部に保持させる保持位置を推定することと、
前記方向ずれが前記所定条件を満たさない場合に、前記取得点と前記保持対象物に関する情報とに基づいて前記保持対象物を前記保持部に保持させる保持位置を推定することと
を含む推定方法。 - 請求項1から12までのいずれか一項に記載の推定装置によって推定した保持位置、又は、請求項13又は14に記載の推定方法を実行することによって推定した保持位置で保持部に保持対象物を保持させる、制御装置。
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US18/833,790 US20240383151A1 (en) | 2022-01-27 | 2023-01-27 | Estimation device, estimation method, and control device |
| EP23747111.5A EP4470731A4 (en) | 2022-01-27 | 2023-01-27 | ESTIMATING APPARATUS, ESTIMATING METHOD AND CONTROL DEVICE |
| CN202380018610.1A CN118591442A (zh) | 2022-01-27 | 2023-01-27 | 推定装置、推定方法以及控制装置 |
| JP2023577043A JPWO2023145901A1 (ja) | 2022-01-27 | 2023-01-27 |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2022011202 | 2022-01-27 | ||
| JP2022-011202 | 2022-01-27 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023145901A1 true WO2023145901A1 (ja) | 2023-08-03 |
Family
ID=87471749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2023/002712 Ceased WO2023145901A1 (ja) | 2022-01-27 | 2023-01-27 | 推定装置、推定方法、及び制御装置 |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20240383151A1 (ja) |
| EP (1) | EP4470731A4 (ja) |
| JP (1) | JPWO2023145901A1 (ja) |
| CN (1) | CN118591442A (ja) |
| WO (1) | WO2023145901A1 (ja) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2024065744A (ja) * | 2022-10-31 | 2024-05-15 | 株式会社東芝 | ハンド、搬送ロボット、制御装置、制御方法、プログラム、及び記憶媒体 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5616478B1 (ja) * | 2013-04-18 | 2014-10-29 | ファナック株式会社 | ワークを搬送するロボットを備えるロボットシステム |
| JP2015073798A (ja) * | 2013-10-10 | 2015-04-20 | 株式会社東芝 | 医用画像診断装置、画像処理装置およびプログラム |
| JP2021010970A (ja) * | 2019-07-05 | 2021-02-04 | 京セラドキュメントソリューションズ株式会社 | ロボットシステム及びロボット制御方法 |
| JP2021035714A (ja) * | 2019-08-22 | 2021-03-04 | オムロン株式会社 | 制御装置、制御方法、及び制御プログラム |
| JP2021131652A (ja) * | 2020-02-19 | 2021-09-09 | 株式会社トプコン | データ構造、記録媒体、プログラム、及びシステム |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6711591B2 (ja) * | 2015-11-06 | 2020-06-17 | キヤノン株式会社 | ロボット制御装置およびロボット制御方法 |
| WO2017083574A1 (en) * | 2015-11-13 | 2017-05-18 | Berkshire Grey Inc. | Sortation systems and methods for providing sortation of a variety of obejcts |
| JP6822929B2 (ja) * | 2017-09-19 | 2021-01-27 | 株式会社東芝 | 情報処理装置、画像認識方法および画像認識プログラム |
| 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 |
| US11833681B2 (en) * | 2018-08-24 | 2023-12-05 | Nvidia Corporation | Robotic control system |
| DE102019122790B4 (de) * | 2018-08-24 | 2021-03-25 | Nvidia Corp. | Robotersteuerungssystem |
| JP6856590B2 (ja) * | 2018-08-31 | 2021-04-07 | ファナック株式会社 | センシングシステム、作業システム、拡張現実画像の表示方法、およびプログラム |
| WO2020184101A1 (ja) * | 2019-03-13 | 2020-09-17 | 日本電気株式会社 | 情報処理装置、駆動制御方法およびプログラム記憶媒体 |
| WO2020212776A1 (en) * | 2019-04-18 | 2020-10-22 | Alma Mater Studiorum - Universita' Di Bologna | Creating training data variability in machine learning for object labelling from images |
| CN114080304B (zh) * | 2019-08-22 | 2024-05-14 | 欧姆龙株式会社 | 控制装置、控制方法及控制程序 |
| US11724401B2 (en) * | 2019-11-13 | 2023-08-15 | Nvidia Corporation | Grasp determination for an object in clutter |
| JP7794641B2 (ja) * | 2020-01-09 | 2026-01-06 | ファナック株式会社 | ワーク搬送システム |
| US12240713B2 (en) * | 2021-03-18 | 2025-03-04 | Dexterity, Inc. | Robotic palletization system with variable conveyor height |
| JP7822205B2 (ja) * | 2022-03-08 | 2026-03-02 | 本田技研工業株式会社 | 遠隔操作システム |
-
2023
- 2023-01-27 WO PCT/JP2023/002712 patent/WO2023145901A1/ja not_active Ceased
- 2023-01-27 JP JP2023577043A patent/JPWO2023145901A1/ja active Pending
- 2023-01-27 US US18/833,790 patent/US20240383151A1/en active Pending
- 2023-01-27 EP EP23747111.5A patent/EP4470731A4/en active Pending
- 2023-01-27 CN CN202380018610.1A patent/CN118591442A/zh active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5616478B1 (ja) * | 2013-04-18 | 2014-10-29 | ファナック株式会社 | ワークを搬送するロボットを備えるロボットシステム |
| JP2015073798A (ja) * | 2013-10-10 | 2015-04-20 | 株式会社東芝 | 医用画像診断装置、画像処理装置およびプログラム |
| JP2021010970A (ja) * | 2019-07-05 | 2021-02-04 | 京セラドキュメントソリューションズ株式会社 | ロボットシステム及びロボット制御方法 |
| JP2021035714A (ja) * | 2019-08-22 | 2021-03-04 | オムロン株式会社 | 制御装置、制御方法、及び制御プログラム |
| JP2021131652A (ja) * | 2020-02-19 | 2021-09-09 | 株式会社トプコン | データ構造、記録媒体、プログラム、及びシステム |
Non-Patent Citations (2)
| Title |
|---|
| "Grasping Detection using Deep Convolutional Neural Network with Graspability", JOURNAL OF THE ROBOTICS SOCIETY OF JAPAN, vol. 36, no. 8, 2018, pages 559 - 566 |
| See also references of EP4470731A4 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118591442A (zh) | 2024-09-03 |
| JPWO2023145901A1 (ja) | 2023-08-03 |
| EP4470731A4 (en) | 2025-12-24 |
| US20240383151A1 (en) | 2024-11-21 |
| EP4470731A1 (en) | 2024-12-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5835926B2 (ja) | 情報処理装置、情報処理装置の制御方法、およびプログラム | |
| JP7111114B2 (ja) | 情報処理装置、情報処理方法及び情報処理システム | |
| CN114494426B (zh) | 用于控制机器人来在不同的方位拿起对象的装置和方法 | |
| US7280687B2 (en) | Device for detecting position/orientation of object | |
| CN107571260B (zh) | 控制机器人抓取物体的方法和装置 | |
| CN107577451B (zh) | 多Kinect人体骨架坐标变换方法及处理设备、可读存储介质 | |
| JP2016099257A (ja) | 情報処理装置及び情報処理方法 | |
| CN114670189B (zh) | 存储介质、以及生成机器人的控制程序的方法及系统 | |
| CN110539299A (zh) | 机器手作业方法、控制器以及机器手系统 | |
| US20230115521A1 (en) | Device and method for training a machine learning model for recognizing an object topology of an object from an image of the object | |
| CN111443703A (zh) | 轨道生成装置、轨道生成方法以及机器人系统 | |
| WO2023145901A1 (ja) | 推定装置、推定方法、及び制御装置 | |
| CN117794709A (zh) | 握持参数估计设备和握持参数估计方法 | |
| KR20230175122A (ko) | 대상물의 조작, 특히 픽업을 위한 로봇 제어 방법 | |
| US20250065495A1 (en) | Holding mode determination device for robot, holding mode determination method for robot, and robot control system | |
| CN115082554A (zh) | 用于控制机器人来拿起对象的装置和方法 | |
| JP7651691B2 (ja) | 保持位置決定装置、及び保持位置決定方法 | |
| JP7801980B2 (ja) | 物体把持方法、プログラム、および物体把持制御装置 | |
| KR20240081551A (ko) | 로봇 제어를 위한 데이터셋 획득 시스템 및 방법 | |
| CN117769483A (zh) | 机器人控制设备、机器人控制系统以及机器人控制方法 | |
| US20250375880A1 (en) | Operation control method, operation control device, and storage medium | |
| US20240351198A1 (en) | Trained model generation method, trained model generation device, trained model, and holding mode inference device | |
| JP7790238B2 (ja) | コンピュータープログラム、生成方法、および、生成装置 | |
| WO2023171687A1 (ja) | ロボット制御装置及びロボット制御方法 | |
| Lee et al. | Balloon burster: A CORBA-based visual servoing for humanoid robot in a distributed environment |
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: 23747111 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202380018610.1 Country of ref document: CN |
|
| ENP | Entry into the national phase |
Ref document number: 2023577043 Country of ref document: JP Kind code of ref document: A |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 18833790 Country of ref document: US |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2023747111 Country of ref document: EP Effective date: 20240827 |