JPH0736519A - Nearmiss checking method for robot - Google Patents

Abstract

PURPOSE:To increase the calculating speed by means of a simple computer in the nearmiss checking method for robots by returning an arithmetic operation to a known interference check calculation against a model which is magnified by an extent equal to a safe distance. CONSTITUTION:In a step S1, the relationship is previously set among a robot, the movement data D1 and D2 on the objects existing in a movable range of the robot, and a safe distance L. The distance L can evade the nearmiss among these robot and objects. In a step S2, the models M1 and M2 which envelope the robot and the objects are produced. In a step S3, both data D1 and D2 are inputted. In a step S4, the distance L is calculated based on the input data D1 and D2 and the relationship X and also the sizes of the robot and the objects are magnified by an extent equal to the distance L. In a step S5, the interference check is carried out in the sizes of the magnified models M1' and M2' so that the nearmiss is checked among the robot and the objects. Thus it is possible to perform an arithmetic operation at a high speed for the nearmiss check by means of a simple computer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a near-miss check method for a robot, and more particularly to a near-miss check method between three-dimensional models used in, for example, a robot off-line programming system or a CAD system.

[0002]

2. Description of the Related Art In order to teach (program) the operation of a robot, it is usually necessary to operate the teaching operation panel or the like, or a person directly moves the robot to actually move the robot. Teaching and recording the position coordinates and the like to create robot control data (teaching data). However, in the above method of actually moving the robot to create the teaching data, the actual work of the robot must be stopped and the operation is troublesome. Therefore, a method using an offline programming operation using a computer may be used. is there. This off-line programming method is described in a programming language called robot language that describes the movements of a robot, or while operating a three-dimensional model of a robot or the like on a graphic display screen, an operation similar to the operation for a real robot. Is performed to create and record teaching data. By the way, in offline programming,
Since the teaching data is not created by actually operating the robot, it is necessary to confirm the robot operation by performing a motion simulation using this teaching data. Checking the interference between the robot and the work or peripheral equipment in this operation simulation is called interference check. This confirmation work is performed by drawing and displaying the robot and the work on the graphic display screen and based on the teaching data. It is troublesome to draw the pictures in time series and visually perform them. For this reason, it is widely practiced to use a three-dimensional mathematical model of a robot or a work to mathematically calculate whether or not there is interference between the models of the robot and the work, as well as the time-series drawing described above. Since such interference detection calculation takes a very long time when the model becomes complicated, various measures have been taken to speed up the calculation. In the operation confirmation of the robot, it may be required not only to detect the interference between the robot and the work, but also to detect with a warning even when the robot approaches the work or the like too much beyond the limit value. . This proximity check is called a near miss check. Conventionally, the near-miss check is performed by calculating the shortest distance between the robot model and the model such as the work, and determining whether it is smaller than a set limit value.

[0003]

In the conventional near-miss check method for a robot as described above, the shortest distance between the robot model and the model such as the work has to be calculated one by one. It takes time. However, when actually confirming the robot operation by the teaching data by simulation,
It is important to detect when there is no interference but there is a risk. Especially, considering a realistic error factor such as blurring during high-speed operation of the robot, it is considered that the strict interference check is not sufficient, and the near-miss check function is essential. Therefore, it is necessary to perform the near miss check operation at high speed, and for this reason, it has been conventionally necessary to use an engineering workstation or the like having a high operation speed. In order to solve the problems in the conventional techniques, the present invention provides a near-miss check method for a robot that improves the near-miss check method for a robot and can perform near-miss check calculation at high speed using a simple computer. It is intended to be provided.

[0004]

In order to achieve the above object, the present invention relates to the relationship between movement data of a robot and an object within the movable range of the robot and a safety distance for avoiding a near miss between the robot and the object. Is set in advance, a model in which the robot and the object are respectively enveloped is created, the movement data is input, and the robot and / or the safe distance calculated from the input movement data by using the relation Alternatively, the method is configured as a near-miss check method for a robot, in which a near-miss check is performed by enlarging the size of a model of an object and performing interference check with the size of the enlarged model. Furthermore, it is a near-miss check method for a robot, in which the dimensions of the model are expanded in units of surface data forming the model. Furthermore, it is a near-miss check method for a robot that expands the dimensions of the model in units of point data that make up the model. The movement data includes not only the loci of the robot and the object, but also the movement speed and the acceleration / deceleration pattern. Further, the locus of the robot includes work patterns and motion patterns such as swinging the hand or changing the locus according to sensor information.

[0005]

According to the present invention, the relationship between the movement data of the robot and the object within the movable range of the robot and the safety distance for avoiding the near-miss check between the robot and the object is preset before the near-miss check. There is. And
A model is created that envelopes the robot and the object. Upon checking, the above movement data is first input. From the input movement data, the dimensions of the model of the robot and / or the object are expanded by the safe distance determined using the relationship. The near miss check is performed by performing the interference check with the dimension of the enlarged model. In this way, the near-miss check can be performed by performing only the interference check without calculating the shortest distance between the robot model and the object model one by one. Further, the dimensions of the model are enlarged in units of surface data that make up the model. Furthermore, the dimensions of the model are enlarged in units of point data that make up the model. In this way, if the size of the model is expanded in units of surface data or point data, the accuracy of the near miss check can be increased as compared with the case of uniformly expanding all the sizes. As a result, it is possible to obtain a near-miss check method for a robot that can perform near-miss check calculation at high speed using a simple computer.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments embodying the present invention will be described below with reference to the accompanying drawings for the understanding of the present invention. It should be noted that the following embodiments are examples of embodying the present invention, and are not intended to limit the technical scope of the present invention. here,
FIG. 1 is a diagram showing a schematic operation procedure by a near miss check method for a robot according to an embodiment of the present invention, FIG. 2 is a schematic configuration diagram of a robot operation simulation device A, and FIG. 3 is a diagram showing a display example of a robot model. 4 is a diagram showing an example of enlargement of the plane data unit of the model, and FIG. 5 is a diagram showing an example of enlargement of the point data unit of the model. As shown in FIG. 1, the robot near miss checking method according to the present embodiment preliminarily defines a relationship X between movement data D1 and D2 of the robot 1 and the object 2 and a safety distance L for avoiding the near miss between the robot 1 and the object 2. After setting (S1), the envelope models M1 and M2 of the robot 1 and the object 2 are created (S2), and the movement data D1 and D2 are set.
(S3), and the model M is equal to the safe distance L obtained by using the relationship X from the input movement data D1 and D2.
The sizes of 1 and M2 are enlarged (S4), and interference check is performed between the enlarged models M1 'and M2' to perform near miss check (S5). FIG. 2 shows an example of a robot simulation apparatus A for embodying such a near miss check method. As shown in FIG. 2, this device A is mainly a data input device 3 (keyboard, mouse, floppy disk, etc.) for inputting a three-dimensional model of an object 2 which is a work target of the robot 1, and arithmetic processing for performing teaching simulation of the robot 1. Device 4 (personal computer, etc.)
Also, using a local memory (not shown) that stores various data used in the simulation, a near-miss check processor 5 that determines a near-miss between the figure of the robot 1 and the figure of the object 2 and the determination result are displayed. A simulation display device 9 (graphic display or the like) for informing the operator of whether a near miss has occurred or not.

The general operation procedure of the device A is as follows. First, the operator inputs the figure data of the object 2, the tip position of the end effector of the robot 1 and the teaching data such as the work command action at each teaching point by the three-dimensional data input device 3. In the arithmetic processing unit 4, the processor 5 in the motion teaching unit 6 and the motion simulation unit 7
You can communicate with. When the processor 5 determines that there is a near miss, the teaching data is changed by an instruction from the arithmetic processing device 4, and the simulation display device 9 performs screen processing and voice processing.
Then, the created teaching data is format-converted by the data transfer unit 8 of the arithmetic processing unit 4, and then transferred to the robot control panel (not shown). In the processor 5 of the apparatus A, a series of operations as shown in FIG. 1 (steps S1 to S
5) is executed. FIG. 3 shows a simulation display device 9
3 shows an example of a model of the robot 1 taught on, for example, a graphic display screen. FIG. 3A is an example of a model enclosing the robot 1, and FIG. 3B is a simple model created for this near miss check.
(C) is a near-miss check simple model with further safety distance L
It is expanded by the amount. In this case, the enlargement indicates the normal direction of each surface of the model. here,
The relationship X between the movement data D1 and D2 of the robot 1 and the object 2 and the safety distance L will be briefly described. As described above, in confirming the operation of the robot, not only the collision between the robot 1 and the object 2 is detected but also a warning is detected when the robot 1 approaches the object 2 too much beyond the limit value. Is required. This limit value is not always a constant value, but the movement data D such as the trajectory of the robot 1 and the object 2 and the movement speed.
1 and D2. For example, when the trajectory of the robot 1 has a curvature of radius R and the robot 1 passes at a velocity V near a stationary object 2 (movement data D2 in this case is zero data), the robot 1 Is applied with a centrifugal force proportional to V ² / R, and this centrifugal force may cause the robot 1 to shake and approach the object 2. In addition to the blurring of the robot, it is necessary to consider the installation error of the object 2 and the like in the safety distance L. The relationship X between the movement data D1 and D2 and the safety distance L can be obtained in advance by analysis or simulation, and the relationship X can be stored in the local memory of the processor 5. Next, the movement data D1 and D2 are input. Since the movement data D1 and D2 are included in the data used by the movement teaching unit 6 and the movement simulation unit 7 of the device A, this data may be used. Then, the safety distance L with respect to the input movement data D1 and D2 can be obtained using the relation X stored in the local memory. Using this safety distance L, models M1 and M2
To increase the dimensions of.

Here, the direction of expansion of the models M1 and M2 will be described. However, the models M1 and M2 are shown in FIG.
In (a), they are represented by rectangular parallelepipeds b and a, respectively.
In FIG. 4 (a), a rectangular parallelepiped a has surfaces A1, A2, ..., A.
5 is expanded (plane movement calculation) by moving the surface normal vector VA1 to VA5 outward by a safe distance L. It should be noted that, for example, if the surface A6 is a processed surface of the object 2, it is not necessary to perform the interference check or the near miss check, and therefore the enlargement is not performed here. In order to realize this expansion, by storing the data representing the surface of the model in association with the normal vector and whether or not the expansion operation is possible, the expansion operation can be speeded up. Such an example is shown in FIG. The information for enlargement includes a flag (or a normal vector) indicating the direction of the surface (which is the outside of the object), a flag indicating whether or not to enlarge in the direction parallel to the surface at the time of enlargement. FIG. 5 (a) also shows an example of enlargement of the model, but in this case, instead of moving the surface,
This is to enlarge (point movement calculation) by moving each vertex in a required direction. In order to realize this expansion, it is necessary to store in the data representing the vertices of the model whether the expansion direction at the time of the expansion operation is to be carried out in the respective directions of the coordinate axes x, y, z in a coded manner. ， Easy to calculate. Such an example is shown in FIG. As described above, the models M1 ′ and M2 ′ whose dimensions are enlarged are provided.
Near-miss check is performed by performing interference check using the processor 5 in the meantime.

As described above, the calculation time of the near-miss check can be shortened by reducing the calculation of the near-miss check to the well-known interference check calculation for the model expanded by the distance, without calculating the distance each time. Further, when this calculation is performed, a simple model with a small number of surfaces (see FIG. 3B) is used as an envelope model, whereby the calculation time can be further shortened. In this case, since the amount of calculation is reduced, the function of the processor 5 can be included in the arithmetic processing device 4 to simplify the device. Furthermore, the robot 1 can also be modeled as an accessory such as a welding wire or a wire drum. By including in, it is possible to perform near miss check with higher accuracy. Furthermore, by enlarging the model in plane data units or point data units, it is possible to make the model finer than in the case of uniformly enlarging, so that the accuracy of the near miss check can be further improved. As a result, it is possible to obtain a near-miss check method for a robot that can perform near-miss check calculation at high speed using a simple computer (for example, a personal computer that can be used easily). It should be noted that the display of the near miss check model (enlarged size model) is usually difficult to see because it overlaps with the model of the robot 1 (envelope model), so the near miss check model is not displayed. However, it may be possible to display it if necessary. Although the safety distance L is automatically calculated in the above embodiment, a person may set the safety distance L in actual use and use the safety distance L directly to enlarge the dimensions of the models M1 and M2. In the above embodiment, the near miss check was performed between the enlarged models M1 'and M2' obtained by enlarging both the models M1 and M2. However, in actual use, only one of the models is enlarged. May be. Alternatively, for example, only one point or one side of the model may be enlarged and used for the near miss check.

The results of consideration of the present invention will be described below. (1) Regarding the expansion of the model size of the robot and / or object, considering that the approach distance (safety distance) is calculated depending on the distance between the two, calculation of the distance is possible even if either the object side or the robot side is expanded. Is possible. Of course, if the distance is apportioned, both may be expanded at the same time as in the above embodiment. However, in reality, the model configuration is simple, or the model can be calculated in advance by the manufacturer, the part that does not need to be calculated can be set, or the model can be replaced by a simple model.・ It is desirable to expand only one of these criteria. This is because expanding both does not reduce the amount of calculation. Therefore, for example, when only a stationary object is handled by the robot, only the robot model needs to be enlarged. On the other hand, not only robots but also objects
The significance of expanding the side is as follows, for example. When the part of the work is processed by gas cutting etc. and the accuracy is poor Especially if only the part with poor processing accuracy and assembly accuracy can be partially expanded, more accurate (fine) near miss check becomes possible. I can say. When there is an error in the work position deviation, the work can be partially enlarged while taking into account the deviation direction in the coordinate system. As described above, in the present invention, the enlarged model size of the robot and / or the object is used for the near miss check. (2) Safe distance The meaning of the safe distance is as follows, for example. When the trajectory of the robot changes, when the teaching data and the position command data are different, and when the calculated position command data and the actually moving track are different (a) Weaving (oscillating motion of the welding torch), multiple layers If you use the functions built into the controller such as the height operation, the operation may exceed the taught position data.
In addition, the position of the work may change due to the rotation or movement of the positioner that fixes the work, and the control trajectory of the robot may change accordingly.

However, this problem can be avoided by simulating these functions (built in the controller) to generate new operation command data and creating movement data based on the position command data or the like. . (B) Considering the robot (and workpiece) dynamics (dynamic characteristics), the position command value and the actual trajectory of the robot may deviate due to inertial force. This matter can be understood by conducting a simulation considering the model dynamics, but it takes a long time to calculate. Therefore, this problem can be avoided by simplifying this as described in the above embodiment and setting an appropriate relationship between the movement data and the safety distance. When the position of the robot or workpiece cannot be determined For example, when performing so-called sensor feedback control using a welding line tracking sensor, trajectory command data is generated on the spot, and it is possible to check the motion and near misses in advance. Can not. These parts should have a large (statistically) large safety distance. When the degree of risk differs depending on the part The case where the degree of risk differs depending on the part is when there is an important part for which the risk of near miss is particularly reduced, but this can be achieved by keeping a large safety distance around the part. good. From the above, in the present invention, the safety distance is used for the near miss check.

[0012]

Since the robot near miss check method according to the present invention is configured as described above, the near miss check operation can be performed without calculating the distance each time.
The calculation time can be shortened by reducing the known interference check calculation to the model expanded by the distance. Moreover, the calculation time can be further shortened by using a simple model with few planes as an envelope model when performing this calculation. In this case, since the amount of calculation is reduced, the function of the processor can be included in the arithmetic processing unit to simplify the device. In addition, the accessories such as the welding wire and the wire drum of the robot are also included in the model. This makes it possible to perform near miss check with higher accuracy. Furthermore, by expanding the model in plane data units and point data units,
Since it can be made finer than when uniformly expanding,
The accuracy of the near miss check can be further improved. As a result, it is possible to obtain a near-miss check method for a robot capable of performing near-miss check calculation at high speed using a simple computer (for example, a personal computer that can be used easily).

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic operation procedure according to a near miss check method for a robot according to an embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of a robot motion simulation apparatus A.

FIG. 3 is a diagram showing a display example of a robot model.

FIG. 4 is a diagram showing an example of enlargement of a plane data unit of a model.

FIG. 5 is a diagram showing an example of enlargement of a point data unit of a model.

[Explanation of symbols]

 1 ... Robot 2 ... Object D1, D1 '... Movement data L ... Safe distance X ... Relationship M1, M1' ... Robot model M2, M2 '... Object model

Claims (3)

[Claims] 1. A relationship between movement data of a robot and an object within a movable range of the robot and a safety distance for avoiding a near miss between the robot and the object is preset, and the robot and the object are enveloped respectively. Create a model
The movement data is input, the dimension of the robot and / or object model is enlarged by the safety distance obtained from the input movement data by using the above relation, and the interference is checked by the dimension of the enlarged model. A near-miss check method for a robot that performs a near-miss check by performing. 2. The near-miss check method for a robot according to claim 1, wherein the size of the model is expanded in units of surface data forming the model. 3. The near-miss check method for a robot according to claim 1, wherein the size of the model is expanded in units of point data forming the model.