JPH09292908A - Control method for robot - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the number of sampling data, to shorten the time required for loading and saving a program and to increase the number of stored programs by changing a sampling period in accordance with the positioning precision of a robot and preventing a part having linearity from being stored as a sampling point. SOLUTION: The sampling period which can be set/changed is decided for obtaining speed and a present sampling position is corrected in accordance with a model based on speed, a spring coefficient and a damper coefficient, which can be set/changed. A direction cosine is obtained in the position and whether copying linearity is judged to be registered in the program or not by a parameter which can be set/changed. When, the operation speed of the robot is slow, the sampling period is prolonged, and it is shortened when it is fast. Thus, sampling position intervals are generally adjusted and correspondence to the positioning precision of the robot can be realized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control method capable of thinning and correcting sampling data of a robot capable of acquiring a motion trajectory while following a work in a force controllable robot.

[0002]

2. Description of the Related Art In some cases, a robot capable of force control acquires a motion trajectory while following a workpiece, and in such control, the regenerative operation is performed by using all sampled data.

[0003]

However, when all of the sampling data are used, the number of sampling data is large, so that the program size is large and it takes time to load and save the sampling data. Further, there is also a problem that the number of programs that can be stored in the controller of the robot is significantly limited because one program size becomes large due to the size of the sampling data.

In view of the above problems, the present invention reduces the number of sampling position data, reduces the program size, shortens the time required to load and save programs, and reduces the number of programs stored in the control device. The purpose is to provide a robot control method for increasing the number of robots.

[0005]

The present invention which achieves the above-mentioned object is characterized by the following matters specifying the invention. (1) In a robot control method in which a controllable robot acquires a motion trajectory while following a work, a sampling cycle for setting the speed is determined, and the speed and the settable spring coefficient and damper coefficient are set. Based on the model, the current sampling position is corrected, and after determining the direction cosine at this position, it is judged whether or not the linearity of the copying is registered in the program by a parameter whose setting can be changed.

Since the sampling cycle is changed according to the positioning accuracy of the robot and the portion having linearity is not stored as the sampling point, the number of sampling data can be reduced.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION Here, an example of an embodiment of the present invention will be described. It is now assumed that the current sampling position is P _(n) , the previous sampling position is P _(n-1) , and the last-previous sampling position is P _(n-2) . Further, the force deviation when the target force F _ref at the current sampling position and the detection force F are ΔF _(n) . This latter relationship is given by the following expression (1). ΔF = F _ref −F (1) On the other hand, the contact state between the hand effector of the robot and the workpiece is modeled as in the following equation (2). That is, f is a force, k is a spring coefficient including the hand effector and the work, x is a displacement, v
= Dx / dt is the speed, and d is the damper coefficient, the contact state is represented by the force f as shown in the following equation (2). f = k · x + d · v (2) The formula (2) of the force f is expressed by the detection force F with respect to the target force F _ref .
The deviation which is ΔF in the above-mentioned expression (1), that is, the contact state at the current position is not lost.

By solving the equation (2) for the displacement x, the following equation (3) is obtained. x = (f−d · v) / k (3) where T is the sampling period, the velocity v is
It is a quotient obtained by dividing the distance between the current position P _(n) and the previous sampling position P _(n-1) by the sampling cycle T. That is, the speed v can be obtained by the following equation (4). v = (P _(n) -P _(n-1) ) / T (4) Thus, the velocity v is obtained by the equation (4), and the equation (3) is obtained.
The displacement x is obtained by

Here, the sampling period T is a parameter whose setting can be changed. That is, the sampling period can be set to the control period or an integral multiple thereof, for example, and can be changed. The position data sampling (actually encoder data reading) cycle performed to control the robot is always constant, and the sampling for trajectory acquisition of the present invention can be set or changed to this control cycle or an integral multiple thereof. is there. Specifically, if the robot's operation speed is slow, the interval between each sampled position becomes short, and if this interval becomes extremely narrower than the positioning accuracy of the robot, it is not necessary to store all sampling data. . Therefore, when the operation speed of the robot is slow, the sampling cycle is lengthened, and when the operation speed of the robot is fast, the sampling cycle is shortened, so that the sampling position intervals are substantially aligned, and the positioning accuracy of the robot can be dealt with. However, when the sampling position interval is too short depending on the sampling period, the sampling position interval becomes sensitive to an error. Therefore, it is necessary to pay attention to this point when determining the sampling position interval.

Further, as in the above equation (2) or (3), in expressing the modeling of the contact state between the hand effector and the work, the spring coefficient k including the hand effector and the work, and The damper coefficient d can be changed. For example, as shown in a simplified form in FIG. 2, when the arm of the robot is contracted A and when it is expanded B, the rigidity is lower when it is expanded, and the rigidity of the entire robot changes depending on the posture. Therefore, it is possible to perform more accurate control by changing the above k and d in accordance with this change.

In this way, the velocity v can be obtained from the equation (4) with the setting of the sampling period T being changeable, and the displacement x can be obtained from the equation (3) with the spring coefficient k and the damper coefficient d being changeable. . Next, the current sampling position P _(n) is corrected by the following equation (5) based on the obtained x. P _(n) = P _(n) -x (5) This correction of the current sampling position P _(n) is performed when the robot and the workpiece are in contact with each other as shown in FIG. Is slightly bent, and the position P _i should be P _e . This is because the position detector (encoder) is attached to the motor and does not detect the actual position of the hand effector, but detects the rotational position of the motor that drives the joint. . That is, the deflection of the arm is not reflected in the encoder data. Since the contact state between the robot and the work is obtained as a model of the spring coefficient and the damper coefficient as described above,
The amount of deflection can be corrected by this model.
That is, the actual position can be replaced by the correction of P _(n) according to the equation (5).

Thus, the actual sampling position P
_(n)And then the previous sampling position P
_(n-1)And this sampling position P_(n)Direction of movement with
(P_(n)−P _(n-1)), And the sampling position P of the previous two times
_(n-2)And the previous sampling position P_{(n -1)}Direction of movement with
(P_(n-1)−P_(n-2)) And direction cosine dir (n), in other words
For example, the degree of linearity is calculated by the following formula [Equation 1]. That is, the figure
Assuming that the sampling is made with a cross as shown in 4,
If there is (x) position data,
Minutes. Therefore, if the movement direction is almost straight,
It is not necessary to obtain sampling data
All that is required is to find dir (n) as the amount of change.

[Equation 1]

For this direction cosine, the sum dirsum of movement direction changes from m times before to this time n is taken by the following equation [Equation 2].

[Equation 2]

Furthermore, the moving direction from m times before to this time
Sum of changes dirsum average dirave (direction average)
It is obtained by the following equation (5). dirave = dirsum / m (5) As a result, dir (n) = cos θ in [Equation 1] above_nAnd this
The direction cosine is the threshold for both dirave shown in equation (5).
 cosθ_thThe average value dirave is within the threshold dirt (dire
ction threshold) by the following equations (6) and (7)
judge. dir (n)> cos θ_th (6) dirave> dirth (7) In other words, if the direction cosine is within the threshold, it is regarded as a straight line.
The average of the sum of changes in the moving direction is within the threshold.
It can be seen as a straight line as a whole.
In this case, cos θ_th, Dirt is a parameter whose design can be changed.
It is. Therefore, at least Equation (6) or (7)
If either side holds, P_(n)= Q _(j)As position data
Q_(j)Update this Q_(j)As the target position data
Add to the program. In this case, it is stored in memory
Is Q_(j)And the sampling data is on a straight line
At this time, only the start point and end point of this straight line are memorized, and
No data needs to be stored.

FIG. 1 is a flow chart showing the whole of the present embodiment, and it is possible to reduce the stored data by making each processing and judgment. Note that n = n + 1 and j = j +
1 means to increment the counter value by one.

[0016]

As described above, according to the present invention, it is possible to obtain position data by using a sampling period or model whose setting can be changed when sampling position data while performing scanning, and perform sampling efficiently and accurately. In addition, since the storage of position data can be reduced and the program size can be reduced, the time required for loading and saving the program is relatively short, and the program size can be reduced and stored in the robot controller. The number increases.

[Brief description of drawings]

FIG. 1 is a flowchart of an example of an embodiment of the present invention.

FIG. 2 is an explanatory view of the rigidity of the arm.

FIG. 3 is an explanatory diagram of deflection.

FIG. 4 is an explanatory diagram of sampling points and storage points.

[Explanation of symbols]

T sampling period d damper coefficient k spring coefficient dir (n), cos θ _n direction cosine dirave average value

Claims (1)

[Claims] 1. A method for controlling a robot, wherein a force-controllable robot obtains a motion trajectory while following a workpiece, sets a sampling cycle in which setting can be changed in order to obtain a speed, and the speed and a changeable spring coefficient are set. A robot characterized by correcting a current sampling position according to a model based on a damper coefficient, determining a direction cosine at this position, and then determining whether or not to register the linearity of the copying in a program with a parameter whose setting can be changed. Control method.