JPH0772910A - Control method for articulated robot - Google Patents

Abstract

(57) [Abstract] [Purpose] To provide a control method for smoothing the motion of a robot in the vicinity of a singularity without stopping the control of the robot in the vicinity of the singularity. [Configuration] In a given position or orientation command calculated joint angle theta ₂ of the W ₂ calculates the distance from the singular point from the absolute value difference between the joint angle at the singular point of the W _2. When the distance from the singular point is equal to or smaller than a predetermined value, the weighted average of the joint angle θ ₁ of W ₁ calculated previously and the calculated value of this time is set as the corrected joint angle θ _1, and the joint angle θ ₃ is set as described above. θ ₁
The value obtained by subtracting the corrected joint angle θ ₁ from the total value of the presently calculated value of θ ₃ and the presently calculated value of θ ₃ is taken as the corrected joint angle θ ₃ .

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control method for an articulated robot capable of stabilizing the operation of a robot near a singular point.

[0002]

2. Description of the Related Art FIG. 3 shows the structure of a robot wrist with three degrees of freedom.
Shown in. Two pivot axes W _1, and W _3, W _1, W between W _{3 1,} W ₃ of the rotary shaft and the joint W with the rotational axis intersecting the vertical
Composed of _two . Generally, these wrists are 3
The robot is attached to the tip of an arm having a degree of freedom to form a 6-degree-of-freedom articulated robot, but only the wrist is shown here. The control of the robot wrist shown in FIG. 3 will be described below. The control of the posture of the robot is generally performed by the method shown in FIG. That is, the command value of the posture (posture command) generated from the position command generator (not shown)
The inverse kinematics conversion unit 1 performs inverse kinematics calculation to convert the joint angle (θ ₁ , θ ₂ , θ ₃ ) that is the target of the robot, and the joint angle control unit 2 converts the joint angle that is the target of the robot joint. Control so that.

[0003]

In the conventional method, when the posture is changed near the singular point (that is, the point where the joint angle of W ₂ becomes 0), the change amounts of W ₁ and W ₃ are very large. Therefore, when the change amount of the angle exceeds the control range, the control is stopped. For this reason,
When working near a singular point, there are problems that the joint movement amounts of W ₁ and W ₃ become too large and control becomes unstable, or control is frequently stopped. In particular, when a command device such as a joystick is used for remote control, noise in the posture command value makes the operation of the robot near the singular point extremely unstable. On the other hand, the applicant of the present application discloses a technique in which the joint at the base (W _{1 in} this example) is fixed and a singular point is passed by the motion of other joints in Japanese Patent Laid-Open No. 189504/1987. However, there is a disadvantage that the motion becomes distorted because the motion of one joint is fixed. Therefore, it is an object of the present invention to provide a control method that stabilizes the motion of a robot in the vicinity of a singular point without stopping the control of the robot in the vicinity of the singular point.

[0004]

In order to solve the above problems, the present invention has a rotation axis W ₂ perpendicularly intersecting the rotation axes of the _two rotation axes W ₁ and W ₃ with each other. For a robot having joints in which joints are arranged as described above, given a position or posture command, the angles (θ ₁ , θ ₂ ,
A method for controlling a multi-joint robot controlled by inverse kinematics transformation into theta _3), the joint angle of the singular point of the W ₂ and from a given position or orientation command calculated joint angle theta ₂ of the W ₂ and Calculate the distance from the singular point from the absolute difference of
If the distance from the singular point is less than or equal to a predetermined value, then W
With the weighted average of the _first previously calculated values of the joint angles theta ₁ and the current calculated value and the correction joint angle theta _1, the joint angle theta ₃ from the total value of the currently calculated value of the present calculated value and the theta ₃ of the theta ₁ A value obtained by subtracting the corrected joint angle θ ₁ is set as a corrected joint angle θ _3, and the robot is controlled by (corrected joint angle θ ₁ , joint angle θ ₂ , corrected joint angle θ ₃ ). is there.

[0005]

By the action wherein said means, in the case of changing the attitude in the singular point near, since the joint angle change of W ₁ is limited by taking a weighted average with the previous calculated values of W ₁ and W ₃ in the vicinity of the singularity Rapid changes are suppressed, and the robot operation can be stabilized.

[0006]

Embodiments Embodiments in which the present invention is applied to the robot shown in FIG. 3 will be described below. FIG. 1 shows an example of an apparatus for carrying out the method of the present invention. FIG. 1 is an example in which the present invention is applied to the conventional example of FIG. 2 in which inverse kinematics conversion is performed from a posture command value generated by a position command generation unit to obtain a target joint angle for control. 1 and 2 are the same as those in the figure, 3 is a singular point distance calculating means, 4 is a load coefficient calculating means, 5 is a load coefficient part, 6
Is a previous value storage means for θ ₁ , 7 and 8 are addition means, and 9 is a subtraction means. The singular point distance calculating means 3 determines the joint angle θ of W _2.
2 , the distance from the singular point is calculated from the absolute value difference from the joint angle of W _{2 at} the singular point, and when the distance is equal to or less than the predetermined value, the value is transmitted to the load coefficient calculating means 4 in the subsequent stage. As shown in FIG. 4, the weighting factor calculation means 4 is a simple increasing function that becomes 0 when the distance to the singular point is 0 and becomes 1 when the weighted variation region near the set singular point is exceeded. The weighting coefficient α is defined and its value is calculated from the distance from the singular point. Because of above configuration, if the distance from the singular point becomes a predetermined value or less, the joint angle of the W ₁ theta ₁
The weighted average of the previous calculated value and this calculated value of is corrected joint angle θ
₁ and the joint angle θ ₃ is calculated from the sum of the current calculated value of θ _{1 and} the current calculated value of θ _3
The value obtained by subtracting ₁ is set as the corrected joint angle θ _3, and the robot is controlled by (corrected joint angle θ ₁ , joint angle θ ₂ , corrected joint angle θ ₃ ).

As a result, when the posture is changed in the vicinity of the singular point, the amount of change in the joint angles of W ₁ and W ₃ is limited, so that even if there is noise in the posture command value, it is stable. It is possible to do the action. Although the wrist portion has been described above, the same applies to the basic shaft portion for positioning.

[0008]

As described above, according to the present invention, when the robot changes its posture near the singular point, the joint W ₂
Of the joint angle of W ₁ and corrected using the preceding value by the distance from the singular point to be determined from the value, W ₃ using this value
By determining, the amount of change in the joint angle of W ₁ and W ₃ is limited, and the motion of the robot near the singular point can be smoothly stabilized.

[Detailed Description of Drawings]

FIG. 1 is a diagram showing an embodiment of the present invention.

FIG. 2 is a control block diagram in the conventional method.

FIG. 3 is a diagram showing a wrist portion with three degrees of freedom of posture of a robot.

FIG. 4 is a diagram showing an example of a weighting coefficient α.

[Explanation of symbols]

1 inverse kinematics conversion unit 2 joint angle control unit 3 singular point distance calculation means 4 load coefficient calculation unit 5 load coefficient unit 6 θ ₁ previous value storage unit 7 and 8 are addition unit 9 subtraction unit

Claims (2)

[Claims] 1. A robot having a joint part, in which joints are arranged so as to have a rotation axis W ₂ perpendicularly intersecting with the rotation axes of the _two rotation axes W ₁ and W ₃ , is provided. The position or posture command is set to the angle of each joint (θ ₁ , θ ₂ ,
A method for controlling a multi-joint robot controlled by inverse kinematics transformation into theta _3), the joint angle of the singular point of the W ₂ and from a given position or orientation command calculated joint angle theta ₂ of the W ₂ and The distance from the singularity is calculated from the absolute value difference of
With the weighted average of the _first previously calculated values of the joint angles theta ₁ and the current calculated value and the correction joint angle theta _1, the joint angle theta ₃ from the total value of the currently calculated value of the present calculated value and the theta ₃ of the theta ₁ A value obtained by subtracting the corrected joint angle θ ₁ is set as a corrected joint angle θ _3, and the robot is controlled by (corrected joint angle θ ₁ , joint angle θ ₂ , corrected joint angle θ ₃ ). Robot control method. 2. The weighted average for obtaining the corrected joint angle θ ₁ is 0 when the distance to the singular point of the weighting coefficient α is 0, and is 1 when the weighted variation area near the set singularity is exceeded. defined by Do increasing function, joint angle of the W ₁ theta ₁
The previous calculation value of is multiplied by α, and the weighted average of this calculation value is calculated as (1-
The method of controlling an articulated robot according to claim 1, wherein the control is performed by multiplying by α and adding.