JP3507032B2 - Method for deriving machine differences of robots - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine difference deriving method, and more particularly to a robot machine difference deriving method capable of easily estimating a robot machine difference. [0002] Robot mechanisms such as industrial robots and NC
As a means for calculating various parameters in a machine tool, a Denabit-Hartenberg notation (hereinafter referred to as “DH notation”) is conventionally used. This expresses the coordinate axes of adjacent drive systems with four parameters consisting of two translation parameters and two rotation parameters. By the way, in the case of an industrial robot, there is generally a machine difference. When a robot is installed and taught, it is necessary to correct a teaching point for each robot and teach a different position. However, the parameters in the DH notation described above do not include parameters related to machine differences for each robot. Therefore, when the DH notation is applied to industrial robots, a means for estimating machine differences and correcting each robot is required. Such means is disclosed in, for example, JP-A-4-2118
In the method disclosed in Japanese Patent Application No. 06, etc., there are five geometric parameters, a link distance a, a link angle θ, an offset d, a twist angle α, and a Y-axis rotation. The rotation angle β of each was given a machine difference, and the machine difference was estimated. However, in such a method, there are as many parameters including machine differences as the five geometric parameters as many as the number of joint axes.
When actually estimating the parameters of a robot having one joint axis, simultaneous equations of 5 × 6 and at least 30 elements are required. The number of reference points to be set at this time is 30 ÷ 3 in consideration of the existence of three data of x, y, and z per reference point, and at least 10 points are required. That is, it is necessary to prepare at least 10 reference points and teach the reference points as teaching points for each robot, so that the setting work becomes very complicated.
Furthermore, as the number of parameters to be estimated increases, the accuracy of the estimated individual correction amounts becomes lower, and as a result, the accuracy of work as a robot deteriorates. Furthermore, since the obtained position data of the reference point is given by a complicated nonlinear equation of each machine difference parameter, it is very difficult to derive this nonlinear simultaneous equation. The present invention has been made to solve the above-described problems of the prior art, and reduces the number of reference points to be created, that is, reduces the number of teaching points, or the correction amount accuracy is high even at equivalent teaching points. Therefore, an object of the present invention is to provide a method for deriving a machine difference between robots so that the work accuracy of the robot is also guaranteed. In order to achieve the above object, the present invention uses a Denabit-Hartenberg notation that represents a robot mechanism having a plurality of joint axes by various parameters. In modeling, only the machine difference parameter relating to micro-rotation is used as the parameter, and then the robot is actually operated, and at least the individual joint axes of the robot.
Measure the 3D position at the reference points set more than a few ,
Finally, based on the three-dimensional position of the reference points obtained communication
The present invention provides a method for deriving a machine difference of a robot characterized in that a machine linear parameter is calculated by creating a linear equation and solving the simultaneous linear equations. With this configuration, the reference point for calculating the machine difference parameter needs to be at least as many as the number of joint axes of the robot. The number can be reduced. In addition, the creation of simultaneous linear equations and the processing for solving them can be performed off-line, so that the overall processing time can be shortened and the computation load on the processing apparatus can be reduced. DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic diagram showing a robot body 1 having 6-axis joints and a robot control device 2 for controlling the robot body 1. As an actual usage pattern, an end effector provided at the tip of the robot body 1
Tools such as hands and tools will be equipped. FIG. 2 shows a schematic flow relating to a machine difference deriving method for a robot according to an embodiment of the present invention. Here, a machine difference derivation method by modeling machine difference parameters when each joint is made independent will be described with reference to the flowchart of FIG. In the flowchart of FIG. 2, “S *” (* is a number) indicates a step number. First, the joint mechanism is modeled by DH notation from the CAD data of the robot (step 1). In the modeling, only the parameters relating to the minute rotation are used as the parameters of the model. Next, the robot is actually operated, and the three-dimensional position of the tip of the robot is measured at n reference points (step 2). Here, n is the number of joint axes of the robot, and n is 6 in this embodiment. Finally, x, y, z per reference point
Based on the existence of the three data, the 3n simultaneous linear equations are constructed from the obtained three-dimensional positions of the n reference points and solved to obtain 18 (3
X6) machine difference parameter is estimated (step 3). Since the processing in step 3 can be performed offline, the overall processing time can be shortened and the calculation load related to the processing apparatus can be reduced. The flow itself shown in FIG. 2 is the same as that of the prior art. In the present embodiment, however, only the parameter relating to the minute rotation is used as the model parameter in the modeling in step 1. Unlike,
Furthermore, since the number of parameters is smaller than that of the prior art, it is different from the prior art in that the number of measurements in step 2 is small. The basic principle of the present invention in the 6-axis robot as shown in FIG. 1 will be described below. FIG.
These are the schematic diagrams shown about the mechanism constant in DH notation. In FIG. 3, the joint n is connected to the joint n + 1, and these mechanism constants are the inter-joint distances a _n ,
It represents a joint angle θ _n , an offset d _n , and a twist angle α _n , and when the joints are parallel, there is a twist angle β _n . The mechanism constant for one of these joints is expressed by an equation (1). ## EQU1 ## If a component representing a machine difference is added to the equation (1) as a minute rotation parallel progression, the equation (2) can be obtained. ## EQU2 ## In the equation (2), I is a unit matrix. In addition, β, γ, and δ are rotation angles in machine differences between the X axis, the Y axis, and the Z axis in the reference coordinate system set for each joint axis. In this equation (2),
Since the rotation angles β, γ, and δ in the machine difference are very small, they can be defined as in Expression (2 ′). ## EQU3 ## Similarly, it is assumed that the machine difference in the translational portion is very small. Here, formula (2) is simplified for explanation,
When the rotation part and the translation part are described separately, the expression (3) is obtained. ## EQU4 ## However, in this equation (3), An, R
n and Pn are respectively the following equations. ## EQU5 ## Here, when expression (3) is expanded, expression (4)
As shown. ## EQU6 ## From this, the expression (2) is an expression representing a minute rotation matrix, but by expressing it as the expression (4), the machine difference matrix may include a minute rotation and a parallel progression component. Understand. In other words, the machine difference matrix can be represented by three parameters (β, γ, δ) relating to rotation as shown in Equation (2), while including a minute rotation and a translational part. Therefore, these three parameters (β,
By obtaining γ, δ), conversion including a machine difference for one joint can be performed. Further, by using the matrix including the translation machine difference, the same effect as when the machine difference is accurately estimated can be obtained. In general, the machine difference of the robot main body often depends on the rotating part, and a sufficient estimation accuracy can be obtained with a machine difference matrix using these three parameters (β, γ, δ). The mechanism constants for any one joint have been described above. In the case of a six-axis robot, the mechanism constants for the other five joints are defined in the same manner. If so, it can be defined to have 18 mechanism constants. From these mechanism constants, the vector acting on the end effector provided at the tip of the 6-axis robot can be expressed by equation (5). ## EQU7 ## In this equation (5), since the matrix An (n = 0, 1,..., 5) including the machine difference parameter of each joint is modeled independently for each joint, there are six or more reference points. By measuring the position of the tip of the robot, the machine difference parameter can be easily derived simply by obtaining a linear simultaneous linear equation having 18 parameters. According to the present invention, a robot having a plurality of joint axes is provided .
When modeling the bot mechanism using the notation of Denabit-Hartenberg, which is expressed by various parameters, the parameters are only machine difference parameters related to micro-rotation, and then the robot is actually operated to at least the number of joint axes of the robot. the three-dimensional position is measured at the predetermined reference point, the three-dimensional position of the last, resulting the reference point
A systematic linear equation was created based on the above, and a machine difference parameter was calculated by solving the systematic linear equation. In setting the reference points, at least the number of joint axes of the robot suffices, and as a result, the number of reference points to be set can be reduced as compared with the conventional one. Further, since the simultaneous linear equations can be created and solved offline, the entire processing time can be shortened and the calculation load on the processing apparatus can be reduced. Thus, according to the present invention, it is possible to reduce the number of reference points, that is, teaching points to be created for machine difference derivation. In other words, the number of reference points to be created is the same. If so, the accuracy of the correction amount of the present invention is higher than that of the conventional one, so that the present invention is more effective than the prior art in this respect. In addition, as described above, generally, as the number of parameters to be estimated is larger, the estimated individual correction amounts are also less accurate, so the method of the present invention in which the number of parameters to be estimated is smaller than that of the prior art is as follows. This means that the accuracy of the estimated individual correction amounts is higher than that of the prior art, and as a result, there is also an effect that the robot operation accuracy is increased.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram showing a robot body 1 having a 6-axis joint and a robot control device 2 for controlling the robot body 1; FIG. 2 is a flowchart showing a schematic flow regarding a method of deriving a machine difference of a robot according to an embodiment of the present invention. FIG. 3 is a schematic diagram showing mechanism constants in DH notation. [Explanation of symbols] 1 Robot body 2 Robot controller

Claims (1)

(57) [Claims] [Claim 1] When modeling a robot mechanism having a plurality of joint axes by using the notation of Denabit-Hartenberg, which represents various parameters, the difference between the parameters is related to micro-rotation. parameter only, and then actually operating the robot, the three-dimensional position is measured at the reference point set at least more than the number of joint axes of the robot, finally, on the basis of the three-dimensional position obtained the reference point Communicating
Create a standing linear equations, instrumental error derivation method of the robot, characterized in that to calculate the instrumental error parameter by solving該連elevational linear equations.