JP2021088019A - Robot system and method for controlling robot system - Google Patents

Abstract

To provide a parallel link type robot that can grip an object piled in bulk arbitrarily on a three-dimensional space without damaging the object.SOLUTION: A robot system estimates a posture and a size of an object on the basis of a three-dimensional image, and dynamically controls a trajectory of the object from a gripping coordinate of a gripping point toward a disposing coordinate of a disposing point on the basis of an estimated result of the posture and the size of the object.SELECTED DRAWING: Figure 2

Description

The present invention relates to a robot system and a control method for the robot system.

Against the background of labor shortage, automation of manual work is progressing by introducing robots. Until now, automation of routine work such as parts cutting and welding work in factories has progressed. It is expected to expand further in the future and be applied to various industrial fields such as construction sites, distribution warehouses, foods, cosmetics, and medical products.

One example is the automation of picking work of goods in a distribution warehouse. The picking work automation robot acquires a 3D image of the article shelf with a 3D camera, detects the presence or absence of the target product by image recognition, and grasps the coordinates while comparing with the image recognition and the 3D model data of the article registered in advance. Is calculated. The trajectory is calculated for the obtained gripping coordinates, and picking is executed.

The problem in picking work automation is that the picking work throughput by the robot is slow. For example, even in a state-of-the-art picking robot for a distribution warehouse, the throughput is only about half that of a skilled human worker.

One of the reasons why the operating throughput of the robot is slow is that the operating speed of the robot itself is slow. Up to now, vertical articulated robots have been often used to grip articles in arbitrary postures that are piled up in three dimensions.

A vertical articulated robot has a large number of joints and can be in an arbitrary posture in a three-dimensional space, and a vertical articulated robot has been widely adopted for three-dimensional bulk picking applications. On the other hand, the vertical articulated robot drives the joints connected in series and the motors that control the joints one by one, and controls so that the final coordinates are reached. Therefore, the larger the number of joints, the longer the control time. It takes.

Further, in the vertical articulated robot, since each joint is connected in series and interacts with each other, it is necessary to calculate the control value of each joint motor including the interaction of each joint. Generally, there are multiple solutions in these motor control calculations, and it is necessary to select the solution that gives the safest posture for the robot from among them. The operating throughput of the robot was slow because it took time to calculate these multiple solutions and to evaluate the multiple solutions and select the solution with the safest posture.

Instead of the vertical articulated robot, as in Patent Document 1, the robot's arms are increased from one to a plurality of parallel connection types, and instead the number of joints is reduced to one joint per arm to automate high-speed picking work. Technology is being developed.

In a parallel link type robot in which such robots with multiple arms are linked and operated in parallel, the joints are lined up in parallel, so that the joints can be controlled in parallel and each joint motor can be controlled independently. Since it can be calculated without calculating the interaction between joints, the operating speed of the robot alone can be increased several times faster than that of a vertical articulated robot.

Japanese Unexamined Patent Publication No. 2019-058979

The problem of automating the picking work by the link structure robot having a plurality of arms described in Patent Document 1 is that the movable posture is limited. The vertical articulated robot can move in three directions of three-dimensional positions X, Y, and Z, and in rotational postures RX, RY, and RZ for each of the three axes.

On the other hand, the parallel link type robot can move only in three directions of X, Y, and Z and RZ, which is a rotational posture about the Z axis, due to a mechanical restriction in which joint motors are connected in parallel. Therefore, it is not possible to grip the articles that are arbitrarily piled up in three-dimensional space, and the range of application is limited to the box-shaped articles that are stacked flat.

In the method of Patent Document 1, the gripping coordinates for gripping the object are dynamically extracted, but the arrangement coordinates for releasing the object are fixed. For this reason, if the placement coordinates are fixedly set assuming that a relatively large object is placed diagonally, the holding of the object will be released from the sky for a thin object placed flat. become. As a result, the object is hit against the work table on the placement side, and there is a risk of damaging the object.

An object of the present invention is to provide a robot system in a parallel link type robot, and a robot system capable of grasping arbitrarily loosely stacked objects in a three-dimensional space without damaging the objects, and a control method for the robot system.

The robot system of one aspect of the present invention is a robot system using a parallel link robot having an imaging unit and a control unit, and the parallel link robot is three-dimensional by a robot hand provided with a rotatable joint mechanism. The objects stacked in bulk are gripped and moved from the gripping side to the arrangement side, the object is gripped at the gripping point on the gripping side, the object is released at the arrangement point on the arrangement side, and the imaging unit is stacked in bulk. The three-dimensional image of the object is acquired, and the control unit estimates the posture and size of the object based on the three-dimensional image, and based on the estimation result of the posture and the size of the object. It is characterized in that the trajectory of the object from the gripping coordinates of the gripping point to the placement coordinates of the placement point is dynamically controlled.

The control method of the robot system according to one aspect of the present invention uses a parallel link robot that grips three-dimensionally stacked objects by a robot hand provided with a rotatable joint mechanism and moves them from the gripping side to the placement side. It is a control method of a robot system, in which a three-dimensional image of the objects stacked in bulk is acquired, the posture and size of the object are estimated based on the three-dimensional image, and on the gripping side, the said at a gripping point. The object is gripped, and on the placement side, the object is released at the placement point, and the object moves from the gripping coordinates of the gripping point to the placement coordinates of the placement point based on the estimation result of the posture and size of the object. It is characterized by dynamically controlling the orbit.

According to one aspect of the present invention, in the parallel link type robot, it is possible to grip an object that is arbitrarily piled up in a three-dimensional space without damaging the object.

It is a figure which shows the robot system of the related technology. It is a figure which shows the robot system of Example 1. FIG. It is a figure which shows the recognition and control processing of a recognition control part. It is a figure which showed the angle of view of a 3D camera and the candidate point of the object gripping coordinate. It is a figure which showed the target coordinates and the motion trajectory of a parallel link robot. It is a figure which showed the intermediate coordinates and the motion trajectory of a parallel link robot. It is a flowchart which shows the control method of the robot system of Example 1. FIG. It is a figure which shows the robot system of Example 2.

First, a method of controlling a robot system of a related technique will be described with reference to FIG.
FIG. 1 is a diagram showing the operation of an autonomous grasping / arranging system for three-dimensional bulk objects of a parallel link robot using fixed arrangement coordinates of related technologies.
As shown in FIG. 1, the picking work automation system by the parallel link robot 101 of the related technology is a robot based on the robot hand 103 including the parallel link robot main body 101 and the joint mechanism 102, the three-dimensional camera unit 109, and the three-dimensional camera data. It has a gripping coordinate estimation unit 110 that plans and controls the operation.

In a robot having a limited operating direction such as the parallel link robot 101, objects 111 and 112 stacked in three dimensions can be gripped and aligned. Therefore, a robot hand 103 provided with a joint mechanism 102 that can rotate three-dimensionally is introduced.

The parallel link robot 101 is attached to the gantry 106 and cannot move autonomously. Therefore, the objects 111 and 112 are conveyed to the arrangement side work table 108 by the grip side conveyor 107.

However, in the recognition / control method of the related technique, the gripping coordinates 104 are dynamically extracted by the gripping coordinate estimation unit 110, but the arrangement coordinates 105 are fixed. Therefore, as shown in FIG. 1A, when the arrangement coordinates 105 are fixedly set assuming the case where the relatively large object 112 is arranged diagonally, the thin object 111 placed flat is not subjected to. As shown in FIG. 1 (b), the grip is released from the sky. As a result, the object 111 is hit against the work table 108 on the arrangement side, and the object 11 may be damaged.

Therefore, in the embodiment, in the parallel link type robot, a robot system capable of grasping arbitrarily loosely stacked objects in a three-dimensional space without damaging the objects and a control method of the robot system are provided.

Therefore, in the embodiment, in the parallel link type robot provided with the autonomous control function and the rotatable joint mechanism, the object gripping position and the arrangement position are estimated and estimated based on the three-dimensional image captured by the three-dimensional camera. The robot operation is controlled based on the gripped position.

Hereinafter, examples will be described with reference to the drawings.

The robot system of the first embodiment will be described with reference to FIG.
As shown in FIG. 2, the picking work automation system by the parallel link robot 101 of the first embodiment includes a parallel link robot main body 101, a robot hand 103 provided with a joint mechanism 102, and a three-dimensional camera unit 109, 3 that constitutes an imaging unit. It has a recognition control unit 204 that plans and controls robot movements based on 3D camera data.

The joint mechanism 102 is composed of, for example, a motor type joint mechanism in which a motor is built and actively operates according to an instruction from the recognition control unit 204. The joint mechanism 102 may include a flexible mechanism such as rubber or a ball made of metal or resin, and may be configured by a joint mechanism that passively deforms in response to an external force.

The parallel link robot 101 is attached to the gantry 106 and cannot move autonomously. Therefore, the object 209 is conveyed to the arrangement side work table 108 by the grip side conveyor 107.

The recognition control unit 204 has a three-dimensional object posture estimation unit 205, a three-dimensional object size estimation unit 206, a grip point / arrangement point coordinate calculation unit 207, and an intermediate coordinate / operation speed calculation unit 208.

The three-dimensional object posture estimation unit 205 estimates the posture of the object based on the three-dimensional image acquired from the three-dimensional camera 109. The three-dimensional object size estimation unit 206 estimates the size of the object. The gripping point / placement point coordinate calculation unit 207 calculates the gripping coordinates 201 of the gripping point of the object and the placement coordinates 202 of the placement point based on the estimated posture and size of the object.

On the gripping side, the gripping point / arrangement point coordinate calculation unit 207 calculates the candidate point on the uppermost surface and the widest gripping area among the plurality of predetermined object gripping candidate points as the gripping coordinate 201 of the gripping point. To do. On the placement side, the placement coordinates 202 of the placement points so that the object 209 is not damaged are calculated based on the posture and size of the object 209.

The intermediate coordinate / operation speed calculation unit 208 functions as an intermediate coordinate calculation unit, and is an intermediate point that is a relay point connecting the grip coordinate 201 of the grip point and the placement coordinate 202 of the placement point based on the posture and size of the object 209. Coordinate 203 is dynamically controlled.

Further, the intermediate coordinate / operation speed calculation unit 208 functions as an operation speed calculation unit, and the movement speed of gripping the object 209 and moving it from the gripping side to the arrangement side based on the estimated posture and size of the object 209. Is dynamically controlled.

In this way, the recognition control unit 204 recognizes the object 209 to be gripped and estimates the object posture, thereby releasing the gripping coordinates 201 of the robot hand (suction pad) 103 and the placement coordinates for releasing the suction and arranging the objects in alignment. The robot trajectory 210 and the operating speed from 202 and the gripping coordinate 201 to the arrangement coordinate 202 are dynamically controlled.

As described above, the robot system of the first embodiment is a robot system using a parallel link robot 101 having a three-dimensional camera 109 (imaging unit) and a recognition control unit 204 (control unit). The parallel link robot 101 grips the three-dimensionally bulked objects 209 by the robot hand 103 provided with the rotatable joint mechanism 102 and moves them from the gripping side to the placement side.

On the gripping side, the object 209 is gripped at the gripping coordinates 201 of the gripping point, and on the placement side, the object 209 is released at the placement coordinates 202 of the placement point.

The three-dimensional camera 109 (imaging unit) acquires a three-dimensional image of the bulk objects 209. The recognition control unit 204 (control unit) estimates the posture and size of the object 209 based on the three-dimensional image, and based on the estimation result of the posture and size of the object 209, the placement point is arranged from the grip coordinates 201 of the grip point. Dynamically controls the orbit 210 of the object 209 toward the arrangement coordinates 202 of.

Here, the recognition control unit 204 (control unit) dynamically controls the trajectory 210 of the object 209 by variably controlling the arrangement coordinates of the arrangement points based on the estimation result of the posture and the size of the object 209. ..

The recognition control unit 204 acquires 3D image data from the 3D camera 109 and transmits it to the 3D object posture estimation unit 205. The 3D camera 109 is irradiated with a signal of a specific wavelength such as a stereo camera that estimates the depth from the parallax with two color cameras and infrared light, and measures the round-trip time that is reflected on the surface of the object and returned to the sensor. Examples include a Time of Flight camera that estimates the depth and a monocular camera method that estimates the depth by a method such as deep learning.

The three-dimensional image data output from the three-dimensional camera 109 consists of two data, one is a depth image that aggregates the depth between the object surface and the sensor captured within the angle of view, and the other is acquired at a specific wavelength. Image data (usually an RGB image acquired over a wide visible area).

The parallel link robot 101 is attached to the gantry 106 and cannot move autonomously. Therefore, the object 209 is conveyed to the arrangement side work table 108 by the grip side conveyor 107. Therefore, at the time of determining the gripping coordinates, the timing at which the three-dimensional image data is taken is acquired, and the target coordinates are determined after taking into account the moving distance of the object by the conveyor 107 within the elapsed time by the recognition and control processing.

For example, the recognition control unit 204 transmits a trigger signal that defines the shooting timing to the three-dimensional camera 109. Alternatively, the 3D image data is constantly transmitted from the 3D camera 109 to the recognition control unit 204, and the 3D image data captured at the same timing as the timing defined by the recognition control unit 204 is selected and acquired. In this way, the time difference between the imaged timing and the end of the recognition process can be estimated, and the coordinates after transportation can be estimated.

As described above, in the robot system of the first embodiment, the three-dimensional camera 109 acquires the three-dimensional data of the loosely stacked objects 209 and transmits the data to the recognition control unit 204. The recognition control unit 204 not only estimates the gripping coordinates 201, but also estimates the size of the object and calculates the arrangement coordinates 202, the intermediate coordinates 203, and the operating speed.

Then, based on the 3D image data obtained from the 3D camera 109, the 3D position / orientation of the target object is first estimated. At the same time, the 3D object model is called from the storage area and the size of the 3D object is estimated. Then, among the plurality of object gripping candidate points defined in advance in the three-dimensional object model, the candidate point on the uppermost surface and having a wide gripping area is calculated as the gripping coordinates 201.

Next, the height of the arrangement coordinates 202 such that the object gently contacts the arrangement surface is calculated from the size information of the three-dimensional object and the three-dimensional position and attitude information. Then, the intermediate coordinate 203 connecting the gripping coordinate 201 and the arrangement coordinate 202 is calculated, and the operating speed of the upper limit robot that does not break the gripping object is calculated and output from the size information and the attribute information of the object.

The recognition and control processing of the recognition control unit 204 will be described with reference to FIG.
Of the 3D image data, only the 2D image is input from the 3D camera 109, and the object is identified by the 2D image analysis (step 301). The gripped object in the acquired image is recognized, and the type of each object is estimated.

When a plurality of objects exist, the object existing downstream of the most traveling direction of the belt conveyor 107 is selected as the gripping target. From the type information of the work selected as the gripping target, the three-dimensional object model of the object is called from the storage area, and the three-dimensional position / orientation estimation is performed (step 302). As a result, it is estimated in what position and posture the object to be gripped exists on the belt conveyor 107.

In the gripping point calculation unit step 303, among the gripping coordinates of a plurality of candidates, the gripping candidate point and the gripping area are compared, and the candidate point having the widest gripping area at the candidate point that is the easiest to grip is calculated as the gripping coordinate 201 of the gripping point. To do.

As shown in FIG. 4, a plurality of gripping candidate points 402 to 405 are calculated for the gripping target object within the angle of view 401, and 402 and 403, which are the gripping points of the most downstream object, are the final candidate points. Compare for selection.

As shown in FIG. 4, at the gripping point 402 and the gripping point 403, since the gripping point 402 has a wider and flatter gripping area, the coordinates of the gripping point 402 are used as the gripping coordinates 201 of the final gripping point. Is output. The grippable area is previously input into the three-dimensional model of the object as information. Alternatively, it may be obtained inductively by machine learning using a two-dimensional image.

Since the rotational posture of the object is obtained in the three-dimensional position posture estimation step 302 with respect to the grip coordinates 201 of the grip point, the rotation posture (RZ) of the Z axis is also given and output to the parallel link robot 101.

In the object placement coordinate estimation step 304, the object identification information obtained in the object identification step 301, the gripping coordinates 201 of the object obtained in the three-dimensional position / orientation estimation step 302 and the gripping point calculation step 303, and the rotational posture of the object are used. , Calculate the placement coordinates 202.

As an example, as shown in FIG. 5, by using the size L1 (503) of the object 209, the size L2 (502) of the object, and the gripped coordinates (501: point A) based on the object identification information, If it is a rigid body, the point C (507), which is the point closest to the conveyor among the objects 209, can be calculated.

Since the grip coordinate point A (501) is a point predetermined for the three-dimensional model of the object 209 in advance, the distance L3 (504) from the grip of the object 209 can be obtained. Normally, it is set so that L3 = L1 / 2 so as to be the center of gravity of the object 209. The coordinates of the point B (506), which is the coordinate point for gripping the object 209, can be calculated as (X + L3cosθ, Y, Z-L3sinθ) by using L3 (504) and the three-dimensional rotation angle θ (505) (510). ).

The coordinates of the point C (507) closest to the conveyor surface in the object 209 can be calculated in the same manner, and can be calculated as (X + L3cosθ-L2sinθ, Y, Z-L3sinθ-L2cosθ).

Therefore, the height H (508) of the object 209 to the conveyor surface is H = Z-L3sinθ-L2cosθ. Therefore, by using the calculated height H (508) with respect to the reference coordinates (X', Y', Z') for the arrangement coordinates (point D: 509), the arrangement coordinates (509) are (X', Y). It can be obtained as', Z'-H).

By doing so, the arrangement coordinates are dynamically controlled according to the posture and size of the object, so that the object can be arranged safely without being damaged. The method of calculating the arrangement coordinates is not particularly limited to the above-mentioned method, and may be calculated in combination with a functional machine learning method such as deep learning.

In the relay point calculation step 305, the intermediate coordinates 203 are calculated by inputting the gripping coordinates 201 and the arrangement coordinates 202. As shown in FIG. 6, the intermediate coordinate 203 is set so as to be between the gripping coordinate 201 and the arrangement coordinate 202.

As shown in FIG. 6B, the low intermediate coordinates 602 are set for the low gripping coordinates 601 and the arrangement coordinates 603. Further, the intermediate coordinates 203 and 602 are always set so that the parallel link robot 101 does not block the field of view of the three-dimensional camera 109.

In addition, the range of the maximum value and the minimum value of the coordinates is limited in advance for the purpose of avoiding interference with obstacles. Further, the intermediate coordinates 203 and 602 are not limited to one point, and a plurality of coordinate points may exist.

As described above, in the relay point calculation step 305, the intermediate coordinates 203 and 602, which are relay points connecting the gripping coordinates 201 and 601 and the arrangement coordinates 202 and 603, are dynamically set based on the estimated posture and size of the object. To control.

In the robot operation speed calculation step 306, the transfer speed of the parallel link robot 101 is controlled. By inputting the object identification result and the attitude estimation result, the object that is registered in advance and is difficult to adsorb or heavy, the object with a steep attitude, or the point where the intermediate coordinates 203 are at a high position with respect to the conveyor surface is conveyed. A process is performed to improve the transfer accuracy by slowing down the speed.

In this way, the robot operation speed calculation step 306 dynamically adjusts the moving speed of gripping the object and moving it from the conveyor 107 on the gripping side to the workbench 108 on the placement side based on the estimated posture and size of the object. Control.

In the robot control step 307, the gripping coordinates 201, the posture (RZ), the arrangement coordinates 202, the intermediate coordinates 203, 602, and the transport speed calculated so far are received as input variables, and are set as various variables in the robot control program defined in advance. By substituting, the robot control program is completed and the process is executed.

The control method of the robot system of the first embodiment will be described with reference to FIG. 7.
First, the bulk objects are imaged by the three-dimensional camera 109 to acquire the three-dimensional data (S1).
Next, two-dimensional image analysis such as CNN is performed (S2). Then, the type of the object is output (S3). In this way, the two-dimensional image analysis (S2) recognizes the type of the object, outputs the object type, and stores it in the buffer (S3). As an image recognition technique, methods such as Convolutional Neural Network (CNN) and template matching are common.

Next, the most downstream object is selected as the object to be gripped (S4). Then, a three-dimensional object model is acquired (S5). Then, three-dimensional data analysis (example: ICP method) is performed (S6). In this way, the model data created in advance and stored in the storage area is called (S5), and the three-dimensional data analysis (S6) is performed.

The Italian Closet Point (ICP) method, which repeatedly obtains the degree of correlation with the 3D image data while moving and rotating the model data to obtain the position coordinates and rotation posture with the highest degree of correlation, and recently, candidates for grasping from the 3D data. The rotational posture of the object to be gripped is obtained by a method of directly calculating the coordinate points by deep learning.

Next, the posture of the object is output (S7). Then, the gripping candidate point is output (S8). Next, the widest gripping point in the gripping area is selected (S9). Then, the coordinates 201 of the gripping point are output (S10). At this time, a gripping candidate point defined in advance for the object is also output (S8), the grippable areas are compared (S9), and the gripping point having the widest gripping area is output (S10). Then, the placement coordinates 202 of the placement points are calculated (S11).
Next, the relay point 203 is calculated (S12). Then, the robot operating speed is calculated (S13). Then, the calculated coordinates and the robot command are output (S14). Next, the robot moves and the object is grasped and arranged (S15).

Finally, it is determined whether or not the gripping and placement of the object is completed (S16). As a result of the determination, when the grasping and arranging of the object is completed, the process ends. On the other hand, as a result of the determination, if the gripping and arranging of the object is not completed, the process returns to S1.

The robot system of the second embodiment will be described with reference to FIG.
The robot system of the second embodiment is a system in which the recognition control unit 204 of the robot system of the first embodiment shown in FIG. 2 is separated into a recognition calculation processing device 801 and a robot control controller 802. Since the other configurations are the same as the configurations of the robot system of the first embodiment shown in FIG. 2, the description thereof will be omitted.

When separated as the robot control controller 802, the imaging timing of the three-dimensional camera 109 is transmitted from the parallel link robot 101 to the recognition calculation processing device 801 via an interface (not shown). When the real-time property of the interface is not guaranteed, it is necessary to determine the gripping coordinates 201 in consideration of the variation in the arrival time of the trigger signal.

According to the above embodiment, the parallel link robot can also grasp and align the articles piled up three-dimensionally in bulk. As a result, it is possible to realize bulk picking by a parallel link robot, which has been difficult in the past. Furthermore, the work throughput can be increased as compared with the automation by the vertical articulated robot.

101 Parallel link robot 102 Rotating joint 103 Robot hand 109 3D camera 107 Grip side belt conveyor 108 Placement side workbench 204 Recognition control unit 205 3D object posture estimation unit 206 3D object size estimation unit 207 Grip point / placement point coordinates Calculation unit 208 Intermediate coordinates / operation speed calculation unit 209 Object 801 Recognition calculation processing device 802 Robot control controller

Claims (13)

A robot system that uses a parallel link robot that has an imaging unit and a control unit.
The parallel link robot
A robot hand equipped with a rotatable joint mechanism grips three-dimensionally bulked objects and moves them from the grip side to the placement side.
On the gripping side, the object is gripped at the gripping point, and on the placement side, the object is released at the placement point.
The imaging unit
Acquire a three-dimensional image of the object piled up in bulk,
The control unit
Based on the three-dimensional image, the posture and size of the object are estimated, and
A robot system characterized in that the trajectory of the object from the gripping coordinates of the gripping point to the placement coordinates of the placement point is dynamically controlled based on the estimation result of the posture and the size of the object. The control unit
The robot system according to claim 1, wherein the trajectory of the object is dynamically controlled by variably controlling the arrangement coordinates based on the estimation result of the posture and the size of the object. The imaging unit is composed of a three-dimensional camera.
The control unit
A three-dimensional object posture estimation unit that estimates the posture of the object based on the three-dimensional image acquired from the three-dimensional camera, and a three-dimensional object posture estimation unit.
A three-dimensional object size estimation unit that estimates the size of the object,
A gripping point / placement point coordinate calculation unit that calculates the gripping coordinates and the placement coordinates of the object based on the estimated posture and size of the object.
The robot system according to claim 1, wherein the robot system comprises. The gripping point / arrangement point coordinate calculation unit
On the gripping side, among a plurality of predetermined object gripping candidate points, the candidate point on the uppermost surface and the widest gripping area is calculated as the gripping coordinates.
The robot system according to claim 3, wherein on the arrangement side, the arrangement coordinates so as not to damage the object are calculated based on the estimated posture and size of the object. The control unit
It is characterized by further having an intermediate coordinate calculation unit that dynamically controls intermediate coordinates, which are relay points connecting the gripped coordinates and the arranged coordinates, based on the estimated posture and the size of the object. The robot system according to claim 1. The control unit
It is further characterized by further having an operation speed calculation unit that dynamically controls the movement speed of grasping the object and moving it from the gripping side to the arrangement side based on the estimated posture and the size of the object. The robot system according to claim 1. The joint mechanism is
The robot system according to claim 1, further comprising a motor-type joint mechanism that actively operates according to an instruction from the control unit. The joint mechanism is
The robot system according to claim 1, further comprising a soft joint mechanism that passively deforms in response to an external force. The parallel link robot is mounted on a gantry and
The robot system according to claim 1, wherein the object is conveyed by a conveyor. It is a control method of a robot system using a parallel link robot that grips three-dimensionally bulked objects by a robot hand equipped with a rotatable joint mechanism and moves them from the gripping side to the placement side.
Acquire a three-dimensional image of the object piled up in bulk,
Based on the three-dimensional image, the posture and size of the object are estimated, and
On the gripping side, the object is gripped at the gripping point.
On the placement side, the object is released at the placement point,
A control method for a robot system, which dynamically controls the trajectory of the object from the gripping point to the placement coordinates of the placement point based on the estimation result of the posture and size of the object. The robot system according to claim 10, wherein the trajectory of the object is dynamically controlled by variably controlling the arrangement coordinates based on the estimation result of the posture and the size of the object. Control method. The robot according to claim 10, wherein the intermediate coordinates, which are relay points connecting the gripping coordinates and the arrangement coordinates, are dynamically controlled based on the estimated posture and the size of the object. How to control the system. 10. The aspect of claim 10, wherein the moving speed of gripping the object and moving it from the gripping side to the arrangement side is dynamically controlled based on the estimated posture and the size of the object. How to control the robot system.

Images