WO2022085560A1 - Robot hand control device, and method for controlling robot hand - Google Patents

Abstract

Provided is a robot hand control device that controls a robot hand having a plurality of fingers to grasp an object to be grasped. The robot hand control device identifies the shape and posture of the object, determines an approaching direction of the robot hand on the basis of the identified shape and posture of the object, moves the robot hand in the determined approaching direction to approach the object, and then causes the robot hand to grasp the object.

Description

Robot hand control device and robot hand control method

The present invention relates to a robot hand control device that controls a robot hand that grips an object, and a method for controlling the robot hand.

There is a robot hand that has multiple fingers and can operate the fingers to grasp an object (work). Such robot hands are used for handling parts in factories and the like.

The robot hand described above is generally used for manipulating an object whose shape and posture are known. Therefore, when the shape and posture of the object to be gripped are not known in advance, how to grip the object has not been sufficiently studied.

The present invention has been made in consideration of the above circumstances, and one of the objects thereof is to provide a robot hand control device capable of appropriately grasping an unknown object, and a robot hand control method. be.

The robot hand control device according to one aspect of the present invention is a robot hand control device that controls a robot hand having a plurality of fingers to grip an object to be gripped, and specifies the shape and posture of the object. The object specifying unit, the determining unit that determines the approach direction of the robot hand based on the shape and posture of the specified object, and the object by moving the robot hand in the determined approach direction. It is a robot hand control device including a grip control unit that causes the robot hand to grip the object after approaching the object.

The robot hand control method according to one aspect of the present invention is a robot hand control method that includes a plurality of fingers and grips an object to be gripped, and includes a step of specifying the shape and posture of the object. A step of determining the approach direction of the robot hand based on the shape and posture of the specified object, and moving the robot hand in the determined approach direction to bring it closer to the object, and then the above. It is a control method of a robot hand including a step of making a robot hand hold the object.

It is an overall schematic diagram of the robot hand control device which concerns on embodiment of this invention. It is a figure which shows the appearance of the robot hand schematically. It is a functional block diagram of the robot hand control device which concerns on embodiment of this invention. It is a flow diagram which shows an example of the flow of the process which the robot hand control device which concerns on embodiment of this invention executes in the preparatory stage of a gripping process. It is a figure explaining the approach direction. It is a flow diagram which shows an example of the flow of the process which the robot hand control device which concerns on embodiment of this invention executes in a gripping process. It is a figure for demonstrating an example of the control content of a gripping operation using a target value. It is a figure for demonstrating an example of the target value set for each finger part.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is an overall schematic view of the robot hand control device 1 according to the embodiment of the present invention. As shown in the figure, the robot hand control device 1 includes a control unit 10, a camera 20, an arm unit 30, and a hand 40.

The control unit 10 is a computer that controls the operations of the arm unit 30 and the hand 40. The content of the process executed by the control unit 10 will be described later.

The camera 20 is fixed at a position where the hand 40 and an object to be gripped by the hand 40 (hereinafter referred to as an object O) can be photographed. The image captured by the camera 20 is transmitted to the control unit 10.

The arm portion 30 is configured to include a plurality of arms 31 connected to each other. These arms 31 are interlocked and driven according to a control command from the control unit 10. The base end of the arm portion 30 is fixed to a base 32 placed on a floor surface or the like, and a hand 40 is attached to the tip of the arm portion 30. By operating these arms 31, the control unit 10 can move the hand 40 to a desired position.

The hand 40 includes a palm portion 41 and a plurality of finger portions 42. FIG. 2 is a diagram showing the appearance of the hand 40. As shown in the figure, in the present embodiment, the hand 40 has five finger portions 42 of the thumb portion 42a, the index finger portion 42b, the middle finger portion 42c, the ring finger portion 42d, and the little finger portion 42e, similar to the human hand. I have.

The palm portion 41 is rotatably connected to the tip of the arm portion 30 within a predetermined range. Further, a palm sensor 43 is arranged on the surface of the palm portion 41. When the palm sensor 43 comes into contact with any object, the palm sensor 43 detects the contact and notifies the control unit 10 of the detection result. The palm sensor 43 may be various sensors such as a pressure sensor.

The base end of each finger portion 42 is connected to the palm portion 41. Further, each finger portion 42 includes one or more joints including a connecting portion to the palm portion 41, and is configured to be flexible at the position of this joint portion. Each finger 42 drives a joint independently of each other according to a control command from the control 10. When gripping the object O, each finger portion 42 bends inward (the side closer to the surface of the palm portion 41) and its tip is pressed against the object O, so that the hand 40 grips the object O. do.

A finger sensor 44 is arranged at the tip of each finger portion 42. That is, the finger sensors 44a-44e are arranged at the tips of the finger portions 42a-42e. The finger sensor 44 detects the magnitude of the force applied to the tip portion when the tip of the finger portion 42 is pressed against the object, and notifies the control unit 10 of the detection result. It is assumed that the finger sensor 44 is a force sensor capable of not only detecting the contact of some object but also measuring the magnitude of the force received by the contact of the object. Further, the finger sensor 44 may be configured to include not only a force sensor that measures the force applied to each finger portion 42, but also a plurality of sensors such as a pressure sensor that detects contact with an object.

Hereinafter, the functions realized by this robot hand control device 1 will be described. The robot hand control device 1 takes an object O whose position and shape are unknown at the start of control as a gripping target, and performs a gripping operation of grasping and lifting the object O with the hand 40. Hereinafter, a series of steps until the robot hand control device 1 drives the arm portion 30 and the hand 40 to perform the gripping operation of the object O is referred to as a gripping step. FIG. 3 is a functional block diagram showing a function provided by the robot hand control device 1 in order to realize such a gripping process. As shown in FIG. 3, the control unit 10 of the robot hand control device 1 is functionally configured to include an object specifying unit 11, a gripping method determining unit 12, and a gripping control unit 13. These functions are realized by the control unit 10 executing a program prepared in advance.

The object specifying unit 11 identifies the position, shape, and posture of the object O in the real space by analyzing the captured image of the camera 20. Specifically, the camera 20 is installed so that the shooting range includes a place (for example, on a table) where the object O to be gripped is supposed to be placed. The object specifying unit 11 identifies the position, shape, and posture of the object O shown in the image by analyzing the captured image of the camera 20.

The gripping method determining unit 12 determines the gripping method of the object O based on the information on the shape and posture of the object O specified by the object specifying unit 11. Specifically, the gripping method determining unit 12 determines specific control contents and control conditions for the arm unit 30 and the hand 40 when executing a series of gripping steps as the gripping method. In particular, in the present embodiment, the gripping method determining unit 12 determines at least the approach direction of the hand 40 with respect to the object O. The approach direction is a moving direction when the hand 40 is moved so as to approach the object O in the gripping step.

Further, the gripping method determining unit 12 grips a condition for actually driving the finger portion 42 to start the gripping operation of gripping the object O (grasping start condition) and a condition for ending the gripping operation (grasping end condition). It may be determined as part of the method. A specific example of the process executed by the gripping method determining unit 12 will be described later.

The gripping control unit 13 executes the gripping process by actually controlling the arm unit 30 and the hand 40 according to the content of the gripping method determined by the gripping method determining unit 12. Specifically, the gripping control unit 13 brings the hand 40 closer to the object O by moving the hand 40 toward the approach direction determined by the gripping method determining unit 12. After that, the movement in the approach direction is stopped at the timing when the predetermined gripping start condition is satisfied, and each finger portion 42 of the hand 40 is driven to start the gripping operation of gripping the object O. Then, at the timing when the predetermined gripping end condition is satisfied, the driving of each finger portion 42 is stopped, and the gripping operation is completed.

The grip control unit 13 determines the timing to start and the timing to end the grip operation while monitoring the detection results of the palm sensor 43 and the finger sensors 44a-44e. That is, the gripping start condition and the gripping end condition include the conditions that the detection results of the palm sensor 43 and the finger sensors 44a-44e should satisfy. Further, the control content of each finger portion 42 during the gripping operation may also be determined according to the output of each finger sensor 44. Specific examples of these controls will also be described later.

Next, an example of specific control contents when the robot hand control device 1 executes the gripping process will be described. First, the processing contents executed by the control unit 10 in the preparatory stage before starting the gripping process for actually driving the arm unit 30 and the hand 40 will be described with reference to the flow chart of FIG.

In the preparatory stage before starting the gripping process, the object identification unit 11 first identifies the correspondence between the robot coordinate system and the camera coordinate system based on the position of the hand 40 included in the captured image of the camera 20 (. S1). Here, the camera coordinate system is a spatial coordinate system based on the shooting position and shooting direction of the camera 20, and the robot coordinate system is a spatial coordinate system based on the position and orientation of the hand 40. The object specifying unit 11 specifies the position and orientation of the hand 40 in the real space with respect to the shooting position of the camera 20 based on the position and appearance of the hand 40 captured in the image captured by the camera 20. As a result, in the gripping process, the control unit 10 can determine the position coordinates of the robot coordinate system based on the position coordinates of the camera coordinate system and control the position of the hand 40 using the position coordinates. A marker image may be arranged on the surface of the hand 40 in order to specify the position of the hand 40. By detecting this marker image from the captured image, the object specifying unit 11 can accurately identify the position and orientation of the hand 40.

Next, the object identification unit 11 recognizes the object O shown in the captured image by analyzing the captured image of the camera 20 (S2). The recognition of the object O can be realized by a known image processing technique or the like.

In this embodiment, it is assumed that the shape of the object O is not known in advance, but it may be known in advance that it is one of a plurality of candidates. In this case, the control unit 10 may acquire information on features such as the shape and surface hardness of these candidates in advance. In this case, the object identification unit 11 identifies which of the plurality of candidates the object O is based on the recognition result of the object O. By doing so, it is possible to facilitate the subsequent specific processing such as the posture of the object O. In addition, gripping start conditions, gripping end conditions, and the like prepared in advance for each candidate may be adopted as the gripping method.

Subsequently, the object identification unit 11 specifies the position of the object O recognized in S2 in the real space (S3). The position of the object O here is specified as the position coordinates of the camera coordinate system. The object specifying unit 11 may simply specify the center position of the entire shape as the position of the object O, but estimates the position of the center of gravity from the shape and specifies the estimated position of the center of gravity as the position of the object O. You may.

Further, the object identification unit 11 specifies the posture of the object O recognized in S2 (S4). Specifically, the object specifying unit 11 specifies in which direction the longitudinal direction of the object O is facing. The longitudinal direction is a direction in which the length from one end of the object O to the other is longer than the other. For example, in the case of an elongated tubular object, the extending direction of the cylinder is the longitudinal direction. By specifying the longitudinal direction, the object specifying portion 11 will specify in what direction the object O is arranged.

Even if the captured image of the camera 20 is just a flat image, it can be expected that the shape and orientation of the object O can be specified with a certain degree of accuracy by a known algorithm. The object specifying unit 11 may use a machine learning algorithm such as deep learning in order to further improve the accuracy of specifying the shape and posture of the object O. Further, in order to directly specify the three-dimensional shape of the object O, a three-dimensional camera capable of measuring the distance to each point on the surface of the object O may be used as the camera 20.

The gripping method determining unit 12 determines the approach direction of the hand 40 based on the longitudinal direction of the object O specified in S4 (S5). Specifically, the gripping method determining unit 12 determines the approach direction so as to intersect the longitudinal direction of the object O (that is, so that the angle formed with respect to the longitudinal direction becomes larger). When the approach direction is orthogonal to the longitudinal direction, the angle formed by the approach direction and the longitudinal direction is maximized, but the approach direction does not necessarily have to be the direction orthogonal to the longitudinal direction of the object O.

As an example, the gripping method determining unit 12 determines the direction in which the angle formed with respect to the longitudinal direction becomes larger from among a plurality of candidate approach directions as the approach direction. In this case, the candidate approach directions may be, for example, two directions, a vertical direction and a horizontal direction. When the candidate approach directions are two directions, the vertical direction and the horizontal direction, the gripping method determining unit 12 determines the vertical direction if the longitudinal direction of the object O is close to the horizontal direction, and the gripping method determination unit 12 if the longitudinal direction of the object O is close to the vertical direction. Select the horizontal direction as the approach direction. Here, the fact that the longitudinal direction of the object O is close to the horizontal direction means that the width of the object O seen from a given reference axis direction along the horizontal plane with the object O placed on the horizontal plane faces the vertical direction. Corresponds to being greater than the height along. On the contrary, the longitudinal direction of the object O is close to the vertical direction when the height of the object O is larger than the horizontal width when the object O is viewed from a given reference axis direction on the horizontal plane. Therefore, when the candidate approach direction is limited to two directions, the vertical direction and the horizontal direction, the object identification unit 11 does not have to accurately specify the angle with respect to the horizontal plane in the longitudinal direction, but the longitudinal direction is either the vertical direction or the horizontal direction. The gripping method determining unit 12 can determine the approach direction as long as it is specified whether it is close (that is, whether the width or the height of the object O is larger).

FIG. 5 is a diagram showing a specific example of the posture and approach direction of the object O. In this figure, the broken line arrow indicates the longitudinal direction of the object O, and the solid arrow indicates the approach direction. FIG. 5A shows an example in which the longitudinal direction of the object O is the vertical direction. In this case, the approach direction is determined to be the horizontal direction, and the hand 40 approaches the object O from the lateral direction along the horizontal plane in the gripping process. On the other hand, FIG. 5B shows an example in which the longitudinal direction of the object O is the horizontal direction. In this case, the approach direction is determined in the vertical direction, and the hand 40 approaches the object O from above the object O along the vertical direction in the gripping step.

Here, the size along the horizontal direction is defined by the width seen from one reference axis direction. However, the object specifying unit 11 may specify the lateral width seen from each of the plurality of reference axis directions, and the direction in which the maximum value is obtained may be the longitudinal direction.

Here, when the approach direction is determined by an automatic calculation with respect to the longitudinal direction of the specified object O, the gripping method determining unit 12 determines the approach direction as the vertically upward direction instead of the vertically downward direction. In some cases. In such a case, the determined approach direction is reversed and corrected so that it is vertically downward.

Further, when the longitudinal direction of the object O is close to the vertical direction and the approach direction is determined to be the horizontal direction, the direction is temporarily determined as one of the directions parallel to the horizontal plane. However, the approach direction in this case may be any direction parallel to the horizontal plane, and does not necessarily have to be a primary determined direction.

Specifically, the gripping method determining unit 12 may determine the approach direction to be actually adopted based on the positional relationship between the hand 40 and the object O at the time when the approach direction is determined. Due to the structure of the arm portion 30, it is difficult to move the hand 40 in the direction toward the base 32 to bring it closer to the object O. Therefore, the gripping method determining unit 12 may determine the approach direction so as to be closer to the direction from the current position of the hand 40 toward the object O. For example, when the hand 40 is located on the right side of the object O when viewed from the camera 20, the direction toward the left side when viewed from the camera 20 is determined as the approach direction. In this way, the hand 40 can be moved to the approach start position with a relatively short movement distance in the preparation stage described later.

In addition, it may not be clear which direction the longitudinal direction of the object O is, such as when the object O has a shape close to a sphere. In such a case, the gripping method determining unit 12 may determine a predetermined direction (for example, a vertical direction) as an approach direction.

Next, the gripping method determining unit 12 determines the approach start position of the hand 40 based on the approach direction determined in S5 and the position of the object O specified in S3 (S6). The approach start position is a position moved by a predetermined distance from this target position to the opposite side of the approach direction with the position of the object O specified in S3 as the target position.

Further, the gripping method determining unit 12 determines the content of the gripping operation and the gripping end condition based on the shape and size of the object O (S7). Specifically, the gripping method determining unit 12 determines the number of finger portions 42 used for the gripping operation. This is because, depending on the size of the object O, it may not be necessary to use all five finger portions 42, or it may be difficult to use them. Further, as described above, when it is specified which of the plurality of candidates the object O is, the gripping method determining unit 12 uses the information stored in advance for each candidate to perform gripping end conditions and gripping. The content of the operation may be determined. The details of the gripping end condition will be described later. Up to this point, the gripping method determining unit 12 determines the gripping method of the object O, and the preparatory step for starting the gripping process is completed.

Next, an example of the control flow when the gripping control unit 13 controls the arm unit 30 and the hand 40 to execute the actual gripping process will be described with reference to FIG.

First, the grip control unit 13 controls the arm unit 30 to move the hand 40 to the approach start position determined in S6 (S11). After that, the movement of the hand 40 is started in the approach direction determined in S5 (S12). As a result, the hand 40 gradually approaches the target position (position of the object O). When the position of the center of gravity of the object O is estimated in S3, the hand 40 is moved with the position of the center of gravity as the target position.

In this state, the grip control unit 13 monitors the detection results of the palm sensor 43 and the finger sensor 44, and monitors whether or not a given grip start condition is satisfied (S13). The gripping start condition is, for example, that one of the sensors detects the contact of the object. In this case, the object for which the sensor detects contact is not necessarily the object O. For example, when the object O is relatively small and the approach direction is vertically downward, the finger sensor 44 can come into contact with the upper surface of the table on which the object O is placed before the palm sensor 43 comes into contact with the object O. There is sex. In such a case, the grip control unit 13 determines that the grip start condition is satisfied when it detects that any of the finger sensors 44 has come into contact with the object.

While the gripping start condition is not satisfied, the gripping control unit 13 continues the control of moving the hand 40 in the approach direction. Then, when it is determined that the gripping start condition is satisfied, the gripping control unit 13 stops the movement of the hand 40 in the approach direction and starts the gripping operation of the object O (S14). Specifically, the grip control unit 13 corresponds to an operation of bending each finger portion 42 used for gripping and bringing the tip thereof closer to the object O (that is, a movement of a finger when a person grips an object by hand). Operation) is executed. In the present embodiment, the thumb portion 42a is connected to the palm portion 41 so as to face the other finger portions 42, similar to the human hand. Therefore, the thumb portion 42a is always the target of the gripping operation. However, not all of the other four fingers 42b-42e must be used for the gripping operation. Depending on the size of the object O, a minimum of one finger portion 42b-42e and a maximum of all four finger portions 42 are selected as targets for the gripping operation.

After starting the gripping operation, the gripping control unit 13 monitors whether or not the given gripping end condition is satisfied (S15). The gripping end condition is a condition for determining that the hand 40 is in a state where the object O can be grasped and lifted by each finger portion 42 to be gripped. The details of the gripping end condition will be described later.

While the gripping end condition is not satisfied, the gripping control unit 13 continues the control of bending each finger portion 42 that is the target of the gripping operation. Then, when it is determined that the gripping end condition is satisfied, the gripping control unit 13 stops the gripping operation of bending each finger unit 42 (S16). After that, the grip control unit 13 executes a predetermined operation such as moving the hand 40 vertically upward to lift the object O.

Hereinafter, the control contents during the gripping operation and specific examples of the gripping end conditions will be described.

First, as a first example, an example in which the gripping end condition is determined by monitoring the position of each finger portion 42 will be described. In this example, the grip control unit 13 monitors the position of each finger portion 42 to be gripped while bending it with a predetermined strength, and each finger portion 42 does not move any more (that is, the position is changed). The gripping operation ends at the point (when it does not change). For example, when it is assumed that the object O has sufficient rigidity and can be gripped with a certain degree of strength, such a gripping end condition can be adopted. In this example, when a drive mechanism such as an actuator included in each finger portion 42 cannot be bent any more, each finger portion 42 is pressed against the surface of the object O with a strength equal to or higher than a predetermined value. It is assumed that there is. When all the finger portions 42 used for the gripping operation are in such a state, it can be determined that the gripping control unit 13 satisfies the gripping end condition.

As a second example, the grip control unit 13 may execute operation control of each finger unit 42 according to the detection result of the finger sensor 44 arranged on each finger unit 42. In this case, the target value of the detection result of the finger sensor 44 is predetermined for each finger portion 42. This target value is determined by the gripping method determining unit 12 as part of the gripping end condition. The grip control unit 13 controls the output value of the drive mechanism included in the corresponding finger unit 42 so that the detection result of each finger sensor 44 approaches a predetermined target value.

Specifically, if the detected value of the corresponding finger sensor 44 is lower than the target value, the finger portion 42 is further bent and the tip thereof is controlled to be closer to the object O. On the contrary, when the detected value of the finger sensor 44 exceeds the target value, the finger portion 42 is changed from the bent state to the extended state, and the tip thereof is controlled to move away from the object O. By continuously executing such control, the detected value of the finger sensor 44 can be matched with the target value for all the finger portions 42 to be gripped. When the detection value of the finger sensor 44 for all the finger portions 42 to be gripped is in a state of matching the target value within a predetermined range, the gripping control unit 13 determines that the gripping end condition is satisfied.

In this example, the grip control unit 13 moves each finger unit 42 so that the displacement amount becomes larger as the finger unit 42 whose detection result of the corresponding finger sensor 44 is farther from the target value during the execution of the gripping operation. May be controlled. In this way, it is possible to control the finger portion 42 in which the magnitude of the force applied to the finger portion 42 is farther from the target value so that the finger portion 42 approaches the target value more quickly.

FIG. 7 is a graph showing an example of such control contents, and shows the relationship between the detected value of the finger sensor 44 and the driving amount of the finger portion 42. The horizontal axis of this figure shows the detected value of the force sensor, and the vertical axis shows the driving amount of the finger portion 42. In the example of this figure, the positive and negative are reversed across the target value, but the larger the difference between the target value and the detected value, the larger the difference between the target value and the detected value, the larger the difference between the detected value and the detected value. The amount of drive is large.

In the example of this figure, the relationship between the detected value of the finger sensor 44 and the driving amount of the finger portion 42 is defined by a cubic function. Specifically, the driving amount y in this figure is calculated by the following formula using the detected value x.
y = ((x-Ft) / α) ³
Where Ft is the target value and alpha is a given adjustment factor. By calculating the drive amount y so as to be proportional to the cube of the detected value x using the cubic function in this way, the absolute value of the drive amount y increases sharply as the detected value x moves away from the target value Ft. It will be calculated.

The target value of each finger portion 42 in this example may be determined according to the positional relationship of each finger portion 42 and the number of finger portions 42 used for the gripping operation. As described above, the gripping method determining unit 12 determines the number of finger portions 42 used for the gripping operation according to the object O. For example, if the object O is sufficiently large, all five finger portions 42 are used for the gripping operation, and if the object O is small, only the thumb portion 42a and the index finger portion 42b are used for the gripping operation. Then, the target value of each finger portion 42 is determined in consideration of the number of finger portions 42 used for the gripping operation.

Specifically, in the gripping method determining unit 12, the total value of the target values set in the finger portion 42 that grips the object O from one side is set in the finger portion 42 that grips the object O from the opposite side. The target value of each finger portion 42 is determined so as to match the total value of the target values. In the present embodiment, since the object O is gripped with the thumb portion 42a and the other finger portions 42 facing each other, the target value of the thumb portion 42a and the total value of the target values of the other finger portions 42 match. You just have to do it. Further, the gripping method determining unit 12 determines the target value for each finger portion 42 so that the target values of the finger portions 42 that grip the object O from the same side are equal to each other.

As an example, a case where the total value of the target values of the finger portions 42 that grip the object O from the same side is set to 0.5N will be described. In the present embodiment, since the thumb portion 42a by itself grips the object O from one side, 0.5N is always set as the target value. For the other fingers 42, equal target values are set according to the number of fingers used for the gripping operation. For example, when only the index finger portion 42b is used for the gripping operation, the target value is set to 0.5N as in the thumb portion 42a. When the two finger portions 42 of the index finger portion 42b and the middle finger portion 42c are used for the gripping operation, the respective target values are set to 0.25N (= 0.5N / 2). FIG. 8 is a diagram schematically showing the relationship between the target values set for each finger portion 42 in this case. When all four fingers 42 are used for the gripping operation, the target value of each finger 42 is set to 0.125N (= 0.5N / 4). By setting the target value of each finger portion 42 in this way, the robot hand control device 1 can grip and grip the object O from two directions facing each other with the same force.

As described above, according to the robot hand control device 1 according to the present embodiment, the object O can be gripped by an appropriate gripping method according to the shape and posture of the object O.

It should be noted that the embodiments of the present invention are not limited to those described above. For example, the shape and configuration of the arm portion 30 and the hand 40 are only examples, and different configurations may be provided. In particular, the number and arrangement positions of the finger portions 42 included in the hand 40 may be different from those described above.

Further, the procedure of the gripping process is not limited to the one described above, and the execution order of the steps may be changed, and steps may be added or deleted.

1 robot hand control device, 10 control unit, 11 object identification unit, 12 gripping method determination unit, 13 grip control unit, 20 camera, 30 arm unit, 31 arm, 32 base, 40 hand, 41 palm part, 42 fingers Department, 43 palm sensor, 44 finger sensor.

Claims (13)

1. A robot hand control device that controls a robot hand having a plurality of fingers to grip an object to be gripped.
   An object specifying part that specifies the shape and posture of the object, and
   A determination unit that determines the approach direction of the robot hand based on the shape and posture of the specified object, and
   A grip control unit that moves the robot hand in the determined approach direction to bring it closer to the object, and then causes the robot hand to grip the object.
   Robot hand control device including.
2. In the robot hand control device according to claim 1,
   The object identification unit specifies the longitudinal direction of the object and
   The determination unit is a robot hand control device characterized in that the approach direction is determined in a direction intersecting the longitudinal direction.
3. In the robot hand control device according to claim 2,
   The determination unit is a robot hand control device characterized in that the approach direction is determined to be either a direction along a vertical direction or a direction along a horizontal direction according to the longitudinal direction.
4. The robot hand control device according to any one of claims 1 to 3.
   The determination unit is a robot hand control device that further determines the approach direction based on the positional relationship between the object and the robot hand.
5. In the robot hand control device according to any one of claims 1 to 4.
   The determination unit determines the approach start position based on the approach direction and the position of the object.
   The grip control unit is a robot hand control device characterized in that the robot hand is moved to the approach start position and then the robot hand is moved in the approach direction.
6. The robot hand control device according to any one of claims 1 to 5.
   The object identification unit identifies the position of the center of gravity of the object and determines the position of the center of gravity of the object.
   The grip control unit is a robot hand control device characterized in that the robot hand is moved in the approach direction with the center of gravity position as a target position.
7. The robot hand control device according to any one of claims 1 to 6.
   Sensors for detecting the contact of an object are provided on each of the plurality of fingers and the palm of the robot hand.
   The grip control unit is a robot hand control device, characterized in that movement in the approach direction is stopped at a timing when any of the sensors detects contact with an object, and gripping of the object is started.
8. In the robot hand control device according to any one of claims 1 to 7.
   Each of the plurality of fingers is provided with a sensor that measures the force applied to the finger.
   The grip control unit causes the robot hand to grip the object by controlling each of the plurality of fingers so that the measured value of the sensor provided on the finger approaches a given target value. A robot hand control device featuring.
9. In the robot hand control device according to claim 8,
   The robot hand control is characterized in that the determination unit determines the number of fingers used for gripping according to the object, and determines the target value for each of the plurality of fingers according to the determined number of fingers. Device.
10. In the robot hand control device according to claim 9.
    In the determination unit, the total value of the target values for one or more fingers gripping the object from one side is the total value of the target values for one or more fingers gripping the object from the opposite side to the one side. A robot hand control device comprising determining the target value for each of the plurality of fingers so as to match.
11. In the robot hand control device according to claim 10,
    The robot hand control is characterized in that the determination unit determines the target value for each of the plurality of fingers so that the target values for each of the plurality of fingers holding the object from the same side are equal to each other. Device.
12. The robot hand control device according to any one of claims 8 to 11.
    The grip control unit is a robot hand control device characterized in that each of the plurality of fingers is controlled so as to move larger as the difference between the measured value of the sensor and the target value increases.
13. It is a control method of a robot hand that has multiple fingers and grips an object to be gripped.
    Steps to specify the shape and posture of the object,
    A step of determining the approach direction of the robot hand based on the shape and posture of the specified object, and
    A step of moving the robot hand in the determined approach direction to bring it closer to the object, and then causing the robot hand to grip the object.
    How to control the robot hand, including.

Images