JPH04201096A - Decision for interference of robot - Google Patents

Abstract

PURPOSE:To quickly execute the decision of the interference of a robot and the body which composes the environment of the circumference, by equipping the shape data of the body which composes a work environment, making the mutual approached bodies as the same group as well, and using the shape model of the work environment expressed in a hierarchic structure as the shape model of this same group. CONSTITUTION:The number of the bodies of the most significant group is not so many because of the shape model 3 of the work environment being made of a hierarchic structure even in case of the number of the body which composes the circumference environment of a robot becoming more with the circumference work environment becoming complicated, and yet those which interfer with the robot are one part of the bodies only existing at the vicinity of the robot in the most significant group. Therefore, the times executing an interference measuring calculation by a calculation device 7 do not increase so much according to the increase of the number of the body composing the environment. So the calculation time necessary for the interference decision is not increased so much. Also, it can cope with enough even for a dynamic environment as well, by having the shape data on the respective body composing the work environment.

Description

[Detailed Description of the Invention] [Purpose of the Invention (Industrial Application Field) The present invention provides a robot interference determination method for determining interference between a robot and an object existing in a work environment in which the robot works. Regarding.

(Prior technology) Robots include industrial robots that perform assembly and processing work in factories, as well as robots for extreme work that perform tasks in place of humans in environments unsuitable for humans. It has come to be used in various fields.

In order to teach the above-mentioned industrial robots and robots for extreme work, the teaching playback method is mainly used for industrial robots, and remote control by humans is used for robots for extreme work.

However, in the teaching playback method for teaching industrial robots how to move, a human actually moves the robot and then stores and records the movements as sensor information.
Since the robot is being regenerated, the only thing that can be done is to let Rojo Soto do the work while the robot is being taught how to move. In the future, as the demand for high-mix, low-volume production increases, there is a possibility that work stoppage periods associated with such teachings will have a significant impact on factory productivity.

Therefore, efforts are being made to develop an offline teaching system in which the robot and its surrounding environment are created as a shape model inside a computer, and the robot's movements are generated on the computer based on these shape models. It is being said.

In an offline teaching system, the operator programs the robot's motion trajectory and checks whether the programmed trajectory is appropriate and whether it will collide with surrounding objects by displaying the robot's motion on the screen. I can do it.

However, from a 2D screen, it may be difficult to determine whether there is interference between a robot moving in 3D and surrounding objects, and a computer calculates whether or not there is interference based on an internal shape model. It is desirable to be able to judge.

Furthermore, when a robot for extreme work is operated by remote control, there is a possibility that a human error in judgment may cause the robot to perform an operation that causes a zero-point collision with objects in the surrounding environment. In order to prevent accidents caused by such operational errors, it may be possible to use sensors to measure whether there are any objects near the robot or throat.

However, in order to measure whether an object exists near the robot, sensors must be attached to the entire surface of the robot. Therefore, in cases where the environment around the robot is known in advance, such as in a nuclear power plant, a geometric model is created inside the computer to check the interference between each link that makes up the robot and objects in the surrounding environment.・One possible method is to perform calculations.

In this way, the shape models of the robot and the objects that make up the surrounding environment are created inside the computer, and it is determined whether the robot and the surrounding environment will collide, or whether they are not close enough to each other to be likely to collide. It is desired to do so.

However, in the interference check between the robot and the surrounding environment as described above, it is determined whether or not there is interference between all the links making up the robot and all the objects existing in the surrounding environment. Therefore, as the surrounding environment becomes more complex, the amount of calculation required to determine interference becomes extremely large, requiring expensive computers with excellent calculation capabilities, and even then, in some cases, the calculation time is extremely long. It takes.

In order to solve these problems, various studies,
Development is underway. For example, JP-A-1-224811
As disclosed in Japanese Patent Application Laid-Open No. 1-1-73205, the shape of objects constituting the environment around the robot is approximated by a sphere, and the collision between this approximate sphere and the robot is first determined. By performing calculations to determine interference between the actual detailed shape and the robot only when there is interference with an approximate sphere, most of the collision determination calculations can be performed between the sphere and the polyhedron that makes up the robot, which requires less calculation. There is a method to reduce the overall amount of calculation for determining interference with a cylinder.

In addition, as disclosed in Japanese Patent Application Laid-Open No. 1-180602, the amount of calculation can be reduced by projecting the robot and objects in the surrounding environment onto a certain plane and performing calculations to determine interference between the projected two-dimensional figures. There are ways to reduce this.

However, although both of the above methods attempt to reduce the amount of calculation required to determine the presence or absence of interference between a pair of objects, they do However, as the surrounding environment becomes more complex, the amount of calculation increases, and it has not been a practical collision detection method for robots.

On the other hand, the Journal of the Robotics Society of Japan Vo I, 5, No.
In the "high-speed interference check method using Octotree" described in Section 3, the robot's work environment is divided into grids, the presence or absence of objects within each grid is calculated in advance, and By representing the divided workspaces in a hierarchical structure, we are attempting to speed up the detection of interference between the robot and its surrounding work environment.

However, this method requires a large amount of memory to represent the entire work environment in a grid. Furthermore, if the work environment changes due to reasons such as the robot moving parts during work, it will be necessary to recalculate or modify the environment model divided into a grid that has been calculated in advance. However, the amount of calculation required for this is quite large, so it is not suitable for cases where the work environment changes.

(Problem to be solved by the invention) In this way, the conventional interference detection method requires a large number of calculations in order to detect interference in an environment with practical complexity, so it requires the use of an expensive computer. Moreover, it required a very long calculation time.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a method for determining interference between a robot and the surrounding environment, which can determine interference between a robot and the surrounding environment with a practical amount of calculation even in a complex and dynamic environment.

[Structure of the invention] (Means for solving the problem) In order to solve the above problem, the invention described in claim (1) provides the following:
The shape of the work environment is calculated using a shape model of the work environment that includes shape data of the objects that make up the work environment, and in which objects that make up the work environment that are close to each other are grouped into the same group, and is expressed in a hierarchical structure as a shape model of this same group. The present invention is characterized in that the interference between the model and the robot is calculated, and only when it is determined that there is interference, the interference with a lower model than the model of the higher level representation is calculated.

In the invention of claim (2), the geometric model of the working environment has an area that constitutes the working environment and an area that does not include objects, and it is calculated whether the geometric model of the working environment and the component parts of the robot are included. It is characterized by the fact that it is made to do so.

(Operation) According to the invention as set forth in claim (1) having the above configuration, the interference calculation with the robot/soto is performed in order from the upper representation to the lower representation model of the shape model of the work environment represented in the hierarchical structure.

In this case, even if the surrounding work environment becomes complex and the number of objects that make up the environment around the robot increases, the shape model of the work environment or the hierarchical structure may cause objects in the top group to The number is not very large, and among the top-level groups, only some objects near the robot interfere with the robot, so the number of collision detection calculations is equal to the number of objects that make up the environment. It does not increase much with the increase.

Therefore, the calculation time required for interference determination does not increase significantly. In addition, by providing shape data to each object that makes up the work environment, even if those objects move, it will be part of the shape data of the moved object and the group that includes the moved object. It is only necessary to change the position and orientation data, and there is no need to change the data regarding the object shape (dimensions), so it can fully cope with dynamic environments.

Further, according to the invention described in claim (2), by including a model of an area that does not include an object in the model of the hierarchically structured work environment, the component parts of the robot are included in the area where no object exists. The calculation means calculates whether the

In the case of normal robots, there are many work environments where a wide operating space is secured around the robot and there are no objects that could become obstacles, and whether each link of the robot is included in the space without obstacles. First, calculate the judgment. This makes it clear that there will be no interference with surrounding objects, eliminating the need to calculate interference with the numerous objects that make up the environment.

Therefore, only the number of collision determinations between the link at the tip of the robot that accesses surrounding objects and the surrounding objects is required, and calculation time can be reduced.

(Example) Next, an example of the robot interference determination method according to the present invention will be described.

FIRST EMBODIMENT As shown in FIG. 1, a collision determination device 1 for a robot 11 (see FIG. 2) includes an environment shape model 3 of the environment in which the robot 11 works, a robot shape model 5, and a robot 11 (see FIG. 2). It is comprised of an interference calculation device 7 that calculates the interference between the robot and the surrounding work environment. More robot 1
A robot shape model changing device 15 is provided that changes the value of the coordinate transformation matrix of each link 13 (see FIG. 2) from one joint angle signal.

Hereinafter, a description will be given of interference determination between a robot and its surrounding environment performed using the robot interference determination apparatus 1 having the above configuration.

Environmental Shape Model 3 Figure 2 shows a computer model 9 of the Japanese Experimental Module at the Space Station FREEDOM. This model 9 includes a robot 11 and the surrounding environment, including a pressurized part where a person lives inside, and an exposed part in which various experimental devices are arranged. The robot 11 in this model is a 6-degree-of-freedom articulated robot, and the number of objects making up the surrounding environment is 44. Note that the objects constituting the environment and the links 13 of the robot 11 are all expressed by combinations of convex stripes and convex stripes.

As shown below, the environment shape model 3 has a hierarchical structure as shown in FIG. 4, by grouping objects constituting the surrounding environment into groups that are close to each other in position.

The above 44 objects are group a (Gr-a) No. 5
~No, 6 Group b (G r -b) No, 8
~No, 17 Group c (G r-c) No,
18-No, 25 Group d (G r-d)
No, 26-No, 31 group e (G
r -e) No, 32~No, 34 group f (Gr -f) No, 36, No, 38~N
o, 48.

In addition, each No. 5 to No. 48 and robot 11
Each link 13 holds the following data values as shape model data.

1. Coordinate transformation matrix between the local coordinate system of the polyhedron and the base coordinate system 2. Numbers of the faces that make up the polyhedron 3. Data for each face ・Coordinate value of one point on the face (value in the local coordinate system) ・Outward normal vector of the surface (value in the local coordinate system) Number 4 of the points composing the surface, data of each point, coordinate value in the local coordinate system Robot shape model changing device 15> Robot shape model changing device 15 In this step, the robot joint angle signal is input manually, the value of the coordinate transformation matrix between the local coordinate system of each link 13 of the robot 11 and the work coordinate system fixed to the space base is calculated, and the robot shape model 5 is calculated.
The value of the coordinate transformation matrix of each link 13 is changed.

Interference Calculation Device 7> The interference calculation device 7 uses the robot shape model 5 modified by the robot shape model modification device 15 and the environment shape model 3 to determine whether objects constituting the robot 11 and the surrounding environment are not interfering with each other. judge whether

<Interference determination procedure> The interference determination procedure will be described below with reference to FIG. FIG. 5(a) shows a procedure for determining link interference, and FIG. 5(b) shows a procedure for determining interference with each obstacle.

First, in step 101 "determination of interference between an obstacle and all links 13", steps 103 and 105 are performed for all links until interference between a link and an obstacle is found. Step 103 ``Calculation of collision determination between link and obstacle'' is shown in FIG. 5(b). In the figure, the highest-level obstacle is a group or object located at the highest level in the environmental shape model expressed in a hierarchical structure. Further, step 105 indicates that step 103 "Calculation of interference determination between each link and obstacle" consists of calculation of interference determination between all the highest-level obstacles and each link. Step 107 ``Calculation of interference between each link and each top-level obstacle'' is performed, and if there is no interference, calculation of interference with all obstacles is sequentially performed.

Even if they are interfering, if the obstacle is a group of several objects, the shape model must contain the shape models of all the objects that make up the group, and the shape model of the obstacle is a group of objects. Since the shape includes gaps between objects, it forms a group (includes in the group)
Unless the geometric models of all lower-order obstacles are judged to be interfering with the links, it will not be possible to determine whether there is really interference.

Therefore, when determining interference between link a and the highest obstacle B, the interference between link a and obstacle B is determined, and if there is interference, all lower obstacles of obstacle i and link a
If there is no interference with all lower-level obstacles, it becomes clear that link a and obstacle B are not interfering with each other, so check the interference between the next obstacle and link a. If it is determined that there is no interference with any of the lower-level obstacles, it is determined that link a and obstacle i are not interfering with each other. If there is an object interfering with link a among the lower obstacles, it is determined that link a is interfering with surrounding obstacles, and the interference check ends there.

The determination of interference between link a and lower-level obstacles is shown in FIG. 5(C), in which the "top obstacle" in FIG. 5(b) is changed to an object constituting a croup. In addition, when determining interference with a lower-level obstacle, if there is no lower-level obstacle (in the case of the lowest-level obstacle)
Assume that there is no interference. The above procedure is performed for all links and all objects (obstacles), and it is found that either the link does not interfere with any object, or the link interferes with a certain object.

Next, an example of interference determination for the model shown in FIG. 3 will be described.

First, calculate the interference between group a and link N011,
If there is interference, link No. with objects No. 5 and No. 6 included in group a.

It turns out that 1 is not interfering. If there is interference, an interference calculation is performed between objects No. 5 and No. 5 included in croup a and link No. 31.

Object No. 5 or No. 5 and link No.

1, it is determined that objects around the robot 11 are interfering with each other, so the calculation is completed and the operator is notified of the interference.

If there is no interference with any of them, then link N001
Interference determination calculation between and group b is performed.

If there is no interference, the object number included in group b,
It turns out that No. 8 to No. 17 and link No. 41 do not interfere. If there is interference, the exposure stand No. and g group No. included in group b.

17 and link No. 1 is calculated. In this way, interference determination between all links and all groups is performed.

In this way, by creating a hierarchical structure for the model of the environment surrounding the robot 11, each link of the robot will not interfere with any higher-level groups other than the group located near the tip of the robot, so collision determination calculations can be performed. If the number of times is considerably smaller than the number of objects that make up the surrounding environment (44), and even if the number of objects that make up the environment increases, the number of top-level groups remains the same, perform collision determination calculation. Since the number of times does not increase significantly, the amount of calculation required for interference determination is small, making it possible to set the calculation time to a practical value.

Calculation for determining interference between each link 13 and objects or groups constituting the surrounding environment is performed using all the vertices of one polyhedron or a plane that extends one face of the other polyhedron infinitely between convex striped polyhedrons. If it is on the outside, it is guaranteed that there is no interference, so follow the steps below.

It is checked whether all the vertices of object B are located outside the surface of all the surfaces of link A, and if such a surface of link A exists, it is determined that there is no interference. If such a surface does not exist even after examining all the faces of link A, it is possible to determine whether there exists a face that is located at all vertices or outside of link A for all faces of object B. Find out.

As a result of the above determination, if there is no surface that satisfies the conditions, it is determined that link A and object B interfere. With this kind of procedure, there are cases where it is determined that there is interference even though there is actually no interference, or the determination is made to prevent robot collisions, and the result is on the safer side. One of the characteristics is that the amount of calculation is reduced.

Second Embodiment Next, a second embodiment will be explained with reference to FIGS. 6 and 7. FIG. 6 clearly shows a region 17 that does not contain any objects in the space base model. Further, FIG. 6 shows the procedure of an interference determination method for determining whether a robot link is included in the area 17 that does not include this object.

Regarding the first link 13, it is first determined whether it is included in the area 17 that does not contain any objects shown in FIG. 6, and if it is included, the first link 13 does not collide with surrounding objects. It turns out.

If it is not included, interference calculations are performed between all objects constituting the environment and the first link 13. The second and subsequent links perform calculations to determine interference between all objects that make up the environment from the beginning. As can be seen from FIG. 6, the first link 13 of the robot 11 is almost always included in the region 17 that does not contain an object.

Therefore, by first determining whether the object is included in the region 17 that does not include the object, it is determined that the object is included, and there is no need to perform calculations to determine interference with objects constituting the environment.

[Effects of the Invention] As explained above, according to the robot collision detection device according to the present invention, since the environment shape model has a hierarchical structure, collision detection between the robot and objects constituting the surrounding environment can be performed using a conventional device. It can be executed in a short time compared to , and even if the number of objects that make up the environment increases, the calculation time does not increase significantly, so it has an excellent effect that it can be applied to actual work. .

Furthermore, in the case of normal robots, there are many work environments where a wide operating space is secured around the robot, and there are no objects or obstacles that could pose an obstacle, and all objects and all links are unnecessarily connected to each other. There is no need to perform interference judgment,
It becomes possible to shorten calculation time.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a collision determination device using the robot collision determination method according to the present invention, FIG. 2 is a perspective view showing an example of a model of the robot and the surrounding environment, and FIG. A perspective view showing the state in which objects constituting the surrounding environment are classified into groups, Figure 4 is a diagram showing the hierarchical structure of the environment shape model, and Figure 5 is a diagram showing the procedure for calculating interference between the robot and the surrounding environment. , FIG. 6 is a diagram showing a region that does not contain an object inside, and FIG. 7 is a diagram showing an algorithm for collision determination calculation using a region that does not contain an object inside. Trobot interference determination device 3...Environmental shape model 5...Robot shape model 7...Interference calculation device 11...Robot 13...Link

Claims (2)

[Claims] (1) In a robot collision detection method for determining interference between objects in the work environment and the robot, the method includes shape data of objects that make up the work environment and classifies objects that are close to each other that make up the work environment into the same group. Using the shape model of the work environment expressed in a hierarchical structure as the shape model of the same group, the interference between the shape model of the work environment and the robot is calculated, and only when it is determined that there is interference, the model of the higher level representation is used. A robot interference determination method characterized in that a robot also calculates interference with a lower model. (2) The shape model of the work environment has an area that constitutes the work environment and an area that does not include objects, and it is calculated whether the shape model of the work environment and the component parts of the robot are included. The robot interference determination method according to claim 1, characterized in that: