JPH04333105A - Robot locus control method - Google Patents

Abstract

PURPOSE:To improve both the stability and the follow-up properties of the servo control by calculating a real servo interpolating position from a primray equation. CONSTITUTION:The interpolating position is calculated in two steps between the teaching positions of a robot axis, and the robot axis is reproduced and controlled based on this calculated interpolating position. The interpolating position of the 2nd step is equal to a real servo interpolating position Pi' set at the present time point by averaging the past real servo interpolating positions Pi-1, P1-2... and the present logical servo interpolating position Pi by a simple equation including a specific weight function ai defined as a coefficient as shown in an equation I. The position Pi is set by dividing the interpolating position of the 1st step by a servo interpolating cycle on a segment connecting linearly both interpolating positions. Thus the position Pi' can be calculated in a simple equation without using a high degree curve. Then the calculating time is shortened owing to reduction of the degree of an arithmetic formula and the interpolating frequency is increased. As a result, the control performance is improved for the locus of a robot.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling the trajectory of a robot, and is also applicable to manipulators, machine tools, transportation machines, etc.

0002

2. Description of the Related Art FIG. 3 shows an example of a device configuration used in a conventional robot trajectory control method. The device shown in the figure is
It is composed of a CPU 1, a memory 2, an actuator 3, and a robot axis 4. Here, the CPU 1 controls the robot axis 4 via the actuator 3 based on the taught position pN of the robot stored in the memory 2 and position information from the robot axis 4.

Further, the memory 2 holds information on the teaching position at which the robot axis should operate, and furthermore, the actuator 3 controls the robot axis 4 according to a command from the CPU 1.
It is what drives the. Therefore, the robot's teaching position p
N-1, pN, pN+1... are stored in the memory 2 in advance, and then, when the robot is to be regenerated, the CPU 1 determines the taught position pN from the memory 2 and the robot position information from the robot axis 4 from the actuator 3. calculate the control signal for the actuator 3 and send the result to the actuator 3.
Output to.

As a result, the robot axis 4 is accurately reproduced on the trajectory defined by the positions pN-1, pN, pN+1, . . . taught in the memory 2.

[0005]

[Problem to be Solved by the Invention] In the conventional robot control method, the taught position pN is stored discretely both in position and time, so when reproducing the path, trajectory errors are functionally ignored. It is necessary to subdivide the line into minute straight lines. If you don't do that,
It is necessary to use a high-speed and expensive CPU.

For this reason, conventionally, the teaching position pN is not directly used for control, but the trajectory interpolation position pj is calculated from the teaching position pN as shown in FIG. I was calculating. Here, the trajectory interpolation position pj is an interpolation point that connects the teaching position with a quadratic or cubic curve and is determined at a time interval such that the trajectory error can be ignored, and usually has an interval of several tens of milliseconds. .

[0007] Furthermore, the servo interpolation position pi is an interpolation point determined to be smooth and accurate within the locus interpolation points pj-1, pj, pj+1, etc. within the servo period, and usually several tens of milliseconds. It has the following spacing: However, while various methods have been proposed for calculating the trajectory interpolation position pj and there is plenty of time for calculation, the calculation method for the servo interpolation position pi is based on the balance between stability of servo control and followability. , was not completely satisfactory.

This is because the servo interpolation position pi needs to be determined in a short period due to constraints on the stability of servo control, but in order to operate smoothly and with good followability, it is necessary to perform high-order curve interpolation. This is because there is no The present invention has been made in view of the above-mentioned prior art, and an object of the present invention is to provide a robot trajectory control method that can achieve both stability and followability of servo control.

[0009]

[Means for Solving the Problems] A feature of the present invention is that when controlling the trajectory of a robot axis, servo interpolation point calculation is performed on both the theoretical servo interpolation position at a certain time point and the actual servo interpolation position at a past time point. On the other hand, by adding and averaging a specific weighting function, calculations can be performed using linear equations without using high-order curves.

[0010] Therefore, during reproduction, it is possible to perform reproduction toward the servo interpolation point resulting from the above. in this way,
By lowering the order of the calculation formula, the calculation time can be reduced and the number of interpolations can be increased.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to embodiments shown in the drawings. The device configuration used in one embodiment of the present invention may basically be the same as the conventional one, and for example, the one shown in FIG. 3 can be used. Therefore, this device is composed of a CPU 1, a memory 2, an actuator 3, and a robot axis 4. The actuator 3 controls the robot axis 4, the memory 2 holds information on the teaching position where the robot axis should operate, and the actuator 3 controls the robot axis 4.
The robot axis 4 is driven by a command from U1.

However, the CPU 1 not only has the same functions and operations as conventional ones, but also has the function of executing the calculations of the following equations (1) and (2). That is, the CPU 1 calculates the distance between the trajectory interpolation positions pj and pj+1 as shown in equation (1) below.
It has a function of calculating the theoretical servo interpolation position pi which is connected by a straight line and divided by the servo interpolation period. Pi = Pj +i・(Pj+1 −Pj)/n
...(1) However, n=T/Δt Δt: Servo interpolation period T: Trajectory interpolation calculation period Furthermore, the CPU 1 calculates the theoretical servo interpolation position pi and the actual servo interpolation position p'i at a past time point held in advance. −1 , p′i−2 … Based on the following formula (
2) has a function of calculating the actual servo interpolation position p'i. Pi ′=(ai ・ki +ai−1 ・p′i−1
+ai−2 ・p′i−2 …+ao )/N


...(2) However, N=ai +ai
-1 +...+ao ai: Weighting coefficient In the device having the above configuration, when reproducing the robot,
It is controlled as shown in FIG.

That is, when reproducing the robot, the CPU 1 first determines the taught positions...pN-1, pN, pN+1.
... is read from the memory 2, a curve passing through the three taught positions pN-1, pN, pN+1 is calculated in the trajectory interpolation calculation cycle T, and further, a curve passing through the two taught positions pN, pN-1 is calculated.
Locus interpolation position...pj-1, pj, pj+1 on the condition that there is no problem on the trajectory connecting the distance with a straight line.
... is calculated.

Next, the CPU 1 connects the locus interpolation positions pj and pj+1 with a straight line, as shown in equation (1), and calculates the theoretical servo interpolation position pi divided by the servo interpolation period Δt. Furthermore, the CPU 1 determines the theoretical servo interpolation position pi
Based on the actual servo interpolation positions p'i-1, p'i-2, .

After that, the CPU 1 determines the actual servo interpolation position pi.
' is output to the actuator 3 and then stored in the memory 2. The robot axis 4 performs a regeneration operation toward the actual servo interpolation position pi' under the control of the actuator 3. In this way, the actual servo interpolation positions p'i-1, pi
' is used to control the robot axis 4, as shown in FIG.
In contrast, it can be seen that the robot trajectory draws a smooth curve.

Therefore, in the present invention, since the actual servo interpolation position is calculated by simplifying it using a linear equation, the calculation time of the CPU 1 can be reduced compared to the case of performing high-order curve interpolation as in the past. Therefore, by increasing the number of servo interpolations per unit time, it is possible to achieve control performance equivalent to or better than interpolation with a high-order curve even with a linear equation. In the above embodiment, the CPU 1 uses the reference point i in equation (2).
And the initial value of the weighting function can be read from a setting device (not shown) such as a preset switch that can be connected to the outside.

[0017]

[Effects of the Invention] As described above in detail based on the embodiments, the present invention reduces the calculation time of the CPU by performing a simplified calculation of the actual servo interpolation point using a linear formula during playback. Can be done. Furthermore, the number of servo interpolations per unit time can be increased, and even with a linear equation, it is possible to achieve control performance equivalent to or better than interpolation with a high-order curve. In particular, by proactively using outputs from past time points, acceleration and deceleration control was automatically performed to improve the lifespan of the aircraft. Furthermore, by averaging the actual servo interpolated positions, the stability of the system against disturbances has been improved.

[Brief explanation of drawings]

FIG. 1 is a flowchart of a robot trajectory control method according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing the principle of the present invention.

FIG. 3 is a configuration diagram showing a conventional robot position reproduction mechanism.

FIG. 4 is an explanatory diagram showing the principle of a conventional control method.

[Explanation of symbols]

1 CPU 2 Memory 3 Actuator 4 Robot axis

Claims (1)

[Claims] 1. A robot trajectory control method in which interpolated positions are calculated in two stages between taught positions of robot axes, and the robot axis is regenerated and controlled along the calculated interpolated positions. The ones in the tier are the past actual servo interpolation positions p'i-1, p'i-2 as shown in the formula below.
...and the current actual servo interpolation position p', which is obtained by integrating a specific weighting function ai as a coefficient for the current theoretical servo interpolation position pi and averaging it as a linear equation.
i, where the theoretical servo interpolation position pi is divided into a plurality of parts at the servo interpolation period on a line segment that linearly connects the first stage of the interpolation positions. trajectory control method. Pi ′=(ai ・pi +ai−1 ・p′i−1
+ai−2 ・p′i−2 …+ao )/N
However, N=ai +ai-1 +...+a
o ai: weighting coefficient