JP2022014138A - Automatic operation-generation method and automatic operation-generation system for welding robot - Google Patents

Abstract

To provide an automatic operation-generation method and automatic operation-generation system for a welding robot which can reduce the pre-work of the welding robot and automatically generate operation of the welding robot on site.SOLUTION: An automatic operation-generation method for a welding robot 1 according to the present embodiment comprises: a setting Step 1 of arranging a three-dimensional measurement sensor 4 at a predetermined location; a measurement Step 2 of dividing a weld location into a plurality of measurement regions and measuring the plurality of measurement regions with the three-dimensional measurement sensor 4; a point cloud processing Step3 of generating extracted surfaces from point cloud data that can be recognized as a plane among measured point cloud data, and generating assumed surfaces from an unrecognizable invalid regions where the point cloud data cannot be obtained; a measurement confirmation Step 4 of confirming whether or not all measurements at the weld location have been completed; and a three-dimensional model generation Step 5 of generating a three-dimensional model of the weld location from the extracted surfaces and the assumed surfaces.SELECTED DRAWING: Figure 2

Description

The present invention relates to an automatic motion generation method and an automatic motion generation system for a welding robot, and in particular, an automatic motion generation method and an operation of a welding robot capable of automatically generating a welding motion of a welding robot brought to a welding place on site. Regarding the automatic generation system.

For example, in the construction of a ship, a hull structure composed of a large number of members such as a longi, a color plate, and a stiffener is the target of welding work. In the early stages of the ship construction process, welding work with a relatively simple structure, such as welding a reinforcing material to a panel, is being promoted, and its automation is also being promoted. As the post-process progresses, the structure becomes three-dimensional and large, so the automatic welding equipment fixed to the factory line has to be increased in size, and the initial cost also increases, so equipment is introduced. It will be a difficult situation. Under these circumstances, it has been studied for many years to bring a welding robot that has been made smaller and lighter to a welding site for automatic welding.

For example, in Patent Document 1, a work model is generated from the work shape information of CAD data, a welding model of a basic welding line consisting of mounting lines of mounting members is generated, and a region dividing line that defines an operating range of a welding robot is used. A cell model that divides the work is generated, the presence or absence of interference between the work and the welding robot is checked for each basic welding line, and if interference occurs, a shortened welding line is generated to the extent that interference does not occur, and each area is generated. Discloses a method of determining the welding direction, order, and path of a welding line, and further specifying welding design information to generate an operation program.

Japanese Unexamined Patent Publication No. 2004-1226

However, as in the invention described in Patent Document 1, when CAD data is used, the CAD data of the welding object must be prepared in advance, and there is a problem that the preliminary work takes time. In addition, when the CAD data is modified, the reflection in the operation program is delayed, a structure that is not reflected in the CAD data is attached, or the CAD data and the actual welding object and its peripheral parts are different in shape. It is not uncommon to be there. In that case, there is a problem that a place other than the welding point is welded, or the welding object or its peripheral portion collides with the welding robot.

The present invention has been devised in view of the above problems, and is capable of reducing the pre-work of the welding robot and automatically generating the motion of the welding robot at the site. The purpose is to provide.

According to the present invention, there is a method for automatically generating a welding motion of a welding robot placed at a predetermined welding place, which is a setting step of arranging a 3D measurement sensor at a predetermined place. The welding site is divided into a plurality of measurement areas and measured by the 3D measurement sensor, and an extraction surface is generated from the point group data that can be recognized as a plane among the measured point group data, and the point group data is generated. A point group processing step for generating an assumed surface from an unrecognizable invalid region that cannot be obtained, a three-dimensional model generation step for generating a three-dimensional model of the welding location from the extracted surface and the assumed surface, and the three-dimensional model. Provided is a method for automatically generating a welding robot motion, which comprises a welding motion generating step of generating a welding motion of the welding robot using the welding robot.

The point cloud processing step may be processed for each measurement step.

In the setting step, the plurality of measurement regions may be processed so as to have a region overlapping with an adjacent measurement region.

The point cloud processing step may recognize that the invalid region is formed by a plane perpendicular to the extraction surface and generate the assumed surface.

The three-dimensional model generation step includes an extraction surface synthesis step of converting the extraction surface into a robot coordinate system and synthesizing it, and a hypothetical surface synthesis step of converting the assumption surface into a robot coordinate system and synthesizing the surface. May be good.

The three-dimensional model generation step may include a plate thickness shape confirmation step of processing a flat surface having a width corresponding to the plate thickness of the steel plate as an end face of the steel plate.

The three-dimensional model generation step may include a shadow portion shape restoration step of extracting and restoring a shadow portion on which the extraction surface and the assumed surface are not formed.

The welding motion generation step may include generating the welding motion in consideration of the inclination of the place where the welding robot is arranged.

Further, according to the present invention, it is an motion automatic generation system of a welding robot that automatically generates a welding motion of a welding robot arranged at a predetermined welding place, and the welding place is divided into a plurality of measurement areas. A 3D measurement sensor for measurement and a calculation device for generating a three-dimensional model of the welding location based on the data of the 3D measurement sensor are provided, and the calculation device can recognize the measured point group data as a plane. An extracted surface is generated from various point group data, an assumed surface is generated from an unrecognizable invalid region where point group data cannot be obtained, and a three-dimensional model of the welding location is generated from the extracted surface and the assumed surface. An automatic motion generation system for a welding robot is provided, which is configured to generate a welding motion of a welding robot using a three-dimensional model.

According to the above-mentioned method for automatically generating motions of a welding robot and the system for automatically generating motions according to the present invention, a three-dimensional model of a welding location is generated using a 3D measurement sensor. Therefore, CAD data is used in advance. There is no need to generate the welding motion of the welding robot. Therefore, according to the present invention, it is possible to reduce the pre-work of the welding robot and automatically generate the operation of the welding robot in the field.

It is a perspective view which shows an example of the welding robot used in the motion automatic generation system of the welding robot which concerns on one Embodiment of this invention. It is an overall flow diagram which shows the motion automatic generation method of the welding robot which concerns on one Embodiment of this invention. It is a flow chart which shows the point cloud processing process. It is a flow chart which shows the 3D model generation process. It is a figure which shows an example of the plane shape after a point cloud processing process, (A) is a 1st measurement area, (B) is a 2nd measurement area, (C) is a 3rd measurement area, (D) is a 4th measurement area. A region, (E) indicates a fifth measurement region, and (F) indicates a sixth measurement region. It is a figure which shows an example of the plane shape after a point cloud processing process, (A) shows the 7th measurement area, (B) shows 8th measurement area, (C) shows 9th measurement area. It is an image diagram which shows the extraction surface synthesis process, (A) shows the synthesis method, (B) shows the synthesis result. It is an image diagram which shows the assumption surface synthesis process. It is an image diagram which shows the plate thickness shape confirmation process. It is an image diagram which shows the shadow part shape restoration process.

Hereinafter, embodiments of the present invention will be described with reference to FIGS. 1 to 10. Here, FIG. 1 is a perspective view showing an example of a welding robot used in the motion automatic generation system of the welding robot according to the embodiment of the present invention.

The welding robot 1 shown in FIG. 1 is a portable welding robot provided with a foldable articulated arm 2. The articulated arm 2 includes, for example, a base 21 arranged on a rotary table 2t, an upper arm 22 rotatably connected to the tip of the base 21, and a lower arm rotatably connected to the tip of the upper arm 22. 23, a wrist portion 24 rotatably connected to the tip of the lower arm 23, and a tool portion 25 rotatably connected to the tip of the wrist portion 24.

The upper arm 22, lower arm 23, wrist portion 24, and tool portion 25 are configured to be folded and arranged on the base 21. The rotary table 2t is arranged on the pedestal 3 and is configured to rotate the articulated arm 2 about the center of the Z axis. A welding torch 2w is arranged at the tip of the tool portion 25.

Further, on the front surface of the upper arm 22, a 3D measurement sensor 4 capable of acquiring a three-dimensional shape as point cloud data is arranged. The 3D measurement sensor 4 is, for example, a distance image sensor capable of acquiring a distance image to a welding object. However, the 3D measurement sensor 4 is not limited to the distance image sensor, and may be any sensor that can obtain three-dimensional point group data. By arranging the 3D measurement sensor 4 in the welding robot 1, the 3D measurement sensor 4 can be carried into the welding place together with the welding robot 1. Further, by arranging the 3D measurement sensor 4 on the upper arm 22, the influence of welding spatter and fume can be reduced as compared with the case where the sensor is installed near the tool portion 25.

In the state shown in FIG. 1, the above-mentioned articulated arm 2 is expressed in XYZ three-axis Cartesian coordinates, around the Z axis of the base 21, around the X axis of the upper arm 22, around the X axis of the lower arm 23, and the wrist. It has a total of 5 degrees of freedom around the X-axis of the portion 24 and around the Y-axis of the tool portion 25. The above-mentioned configuration of the articulated arm 2 is merely an example, and is not limited to the configuration shown in the figure.

On the pedestal 3, a control box 5 for accommodating the arithmetic unit of the automatic motion generation system of the welding robot 1 and the control device of the welding robot 1 is arranged. Further, a handle 6 for transporting the welding robot 1 is arranged on the pedestal 3. The handle 6 is rotated above the folded articulated arm 2 during transportation, and is rotated forward of the pedestal 3 as shown at the time of installation.

Further, handles 31 for lifting the welding robot device 1 mounted on the floor may be arranged on both sides of the pedestal 3. Further, a laser pointer 32 for positioning the welding robot device 1 may be arranged in front of the pedestal 3. Further, a fixing magnet (not shown) composed of a permanent magnet or an electromagnet and a leg portion 33 may be arranged on the bottom of the pedestal 3. Further, although not shown, a sensor for measuring the inclination of the place where the welding robot 1 is arranged, such as an accelerometer or an inclinometer, may be arranged on the pedestal 3.

The motion automatic generation system of the welding robot 1 according to the present embodiment is a motion automatic generation system that automatically generates the welding motion of the welding robot 1 arranged at a predetermined welding place, and has a plurality of measurement areas of the welding place. It is equipped with a 3D measurement sensor 4 that measures separately and a calculation device that generates a three-dimensional model of the welding location based on the data of the 3D measurement sensor 4, and the calculation device is a welding robot based on the flow described later. It is configured to automatically generate the welding motion of 1.

Next, a method for automatically generating an operation of the welding robot 1 according to an embodiment of the present invention will be described with reference to FIGS. 2 to 10. FIG. 2 is an overall flow chart showing a method for automatically generating an operation of a welding robot according to an embodiment of the present invention. FIG. 3 is a flow chart showing a point cloud processing process. FIG. 4 is a flow chart showing a three-dimensional model generation process.

The method for automatically generating the motion of the welding robot 1 according to the present embodiment is a method for automatically generating the welding motion of the welding robot 1 arranged at a predetermined welding place, and the 3D measurement sensor 4 is designated. The setting process Step 1 to be placed in the place, the measurement process Step 2 to divide the welding place into a plurality of measurement areas and measure with the 3D measurement sensor 4, and the point group data that can be recognized as a plane among the measured point group data are extracted. A point group processing step Step 3 that generates a surface and generates an assumed surface from an unrecognizable invalid area where point group data cannot be obtained, and a measurement confirmation step Step 4 that confirms whether or not all measurements at the welding site have been completed. A three-dimensional model generation step Step 5 that generates a three-dimensional model of the welding location from the extracted surface and the assumed surface, a work setting process Step 6 that inputs data related to welding such as the welding location and leg length, and a welding robot using the three-dimensional model. A predetermined welding operation generation step Step 7 for generating a welding motion, a construction order setting step Step 8 for setting the construction order of welding points included in the three-dimensional model, and a predetermined welding robot 1 based on the automatically generated welding motion and the construction order. It includes a welding work step Step 9 for executing welding work.

The setting step Step 1 is a step of arranging the 3D measurement sensor 4 in order to measure a predetermined area of the welding place. In the present embodiment, the 3D measurement sensor 4 is set by arranging the welding robot 1 at a predetermined place and adjusting the direction thereof.

The measurement step Step 2 uses, for example, a sensor capable of obtaining a distance image by irradiating a pattern light (infrared ray) such as a random dot from a 3D measurement sensor 4 and photographing it with one or a plurality of infrared cameras. , Is a process of acquiring point group data of a measurement area from the distance image.

The point cloud processing step Step 3 is a step of extracting or generating a planar shape from the point cloud data. Regularly arranged point cloud data is acquired from a plane (for example, the web surface of Longi) which is arranged perpendicular to the floor surface and is located at a position facing the 3D measurement sensor 4. On the other hand, in a horizontal plane such as a stiffener arranged perpendicular to the web surface of Longi, the incident angle of the irradiation light from the 3D measurement sensor 4 becomes shallow, and the point cloud cannot be measured on that plane, and the 3D measurement sensor cannot be measured. A region where the point cloud data is missing is formed from the viewpoint from 4.

In the present embodiment, a plane extracted from the point cloud data that can be recognized as a plane among the point cloud data is defined as an "extracted surface", and an unrecognizable region where the point cloud data cannot be obtained from the viewpoint of the 3D measurement sensor 4. Is defined as an "invalid area". In addition, the plane generated from the invalid area is defined as the "assumed plane".

The point cloud processing step Step 3 includes a process of generating an extraction surface from the point cloud data, and includes a process of recognizing that the invalid region is formed by a plane perpendicular to the extraction surface and generating an assumed surface. .. Specifically, the point cloud processing step Step 3 is processed based on the flow shown in FIG.

As shown in FIG. 3, the point group processing step Step 3 is coordinate-converted by a data acquisition step Step 31 for acquiring distance image data and a coordinate conversion step Step 32 for converting distance image data into three-dimensional coordinates of the sensor coordinate system. From the local plane calculation step Step 33 that calculates the local plane from the data, the labeling step Step 34 that labels the local plane, the shape extraction step Step 35 that extracts the plane shape (extraction surface), and the distance image data from the viewpoint of the 3D measurement sensor 4. It includes an invalid area extraction step Step 36 for extracting an invalid area and an assumed surface generation step Step 37 for generating an assumed surface from the invalid area.

In the process from the coordinate conversion step Step 32 to the shape extraction step Step 35, after processing the data acquired by the 3D measurement sensor 4 such as downsizing, a general method such as normal direction calculation of a minute region and judgment by distance is used. It is possible, and detailed description is omitted here. In the labeling step Step 34, for example, data is divided and labeled for each plane based on the angle formed by the surface normal, the distance between the surfaces, and the like. Further, when the 3D measurement sensor 4 is a 3D sensor capable of directly acquiring point cloud data, the data acquisition step Step 31 and the coordinate conversion step Step 32 are the same process.

The invalid region is composed of a surface where the incident angle of the laser beam is shallow or a region where there is no object in the measurable range of the 3D measurement sensor 4. Therefore, in the invalid region extraction step Step 36, a minute region is ignored and an elongated region (when a relatively elongated member such as a large steel structure is a measurement target) is cut out to reduce the amount of data processing and the invalid region. To sort out. In the invalid area extraction step Step 36, when the 3D measurement sensor 4 is a 3D sensor capable of directly acquiring point cloud data, the data creation step corresponding to the distance image from the viewpoint of the 3D measurement sensor 4 is performed first. You may do it in.

In the assumed surface generation step Step 37, a surface shape candidate orthogonal to the peripheral surface of the invalid region is created with reference to the position of the 3D measurement sensor 4, and a surface shape in which the invalid region is a shadow is generated. For example, when the surface is higher than the viewpoint of the 3D measurement sensor 4, the invalid area has a trapezoidal shape in which the lower base is longer than the upper base, and when the surface is lower than the viewpoint of the 3D measurement sensor 4, the invalid area is. It has a trapezoidal shape with an upper base longer than the lower base.

The measurement confirmation step Step 4 is a step of confirming whether or not the measurement of the entire welding location has been completed (whether or not the measurement of a predetermined number of times has been completed). If all the measurements have not been completed (N), the process returns to Step 1, and if all the measurements have been completed (Y), the process proceeds to the next process.

In the setting step Step 1, the welding robot 1 is rearranged in a predetermined place, or the articulated arm 2 is moved to change the direction and posture of the 3D measurement sensor 4 to set the 3D measurement sensor 4. At this time, the 3D measurement sensor 4 is set so that the plurality of measurement areas have an area overlapping with the adjacent measurement area. By such processing, point cloud data of the same surface can be acquired in a plurality of measurement regions, recognition of the same surface can be easily performed, and plane composition can be easily processed.

Here, FIG. 5 is a diagram showing an example of the planar shape after the point cloud processing step, where (A) is the first measurement region, (B) is the second measurement region, and (C) is the third measurement region. (D) shows the fourth measurement area, (E) shows the fifth measurement area, and (F) shows the sixth measurement area. 6A and 6B are views showing an example of a planar shape after the point cloud processing step, in which FIG. 6A shows a seventh measurement area, FIG. 6B shows an eighth measurement area, and FIG. 6C shows a ninth measurement area. There is.

The planar shapes shown in FIGS. 5 (A) to 6 (C) indicate the planar shapes extracted or generated based on the point cloud data of the measurement area measured at the Nth time (N is an integer of 1 to 9). ing. In the planar shapes of the first measurement region to the sixth measurement region shown in FIGS. 5A to 5F, the invalid region is not extracted and is composed only of the extraction surface. Further, the planar shape shown in the right figure of FIGS. 6 (A) to 6 (C) includes the assumed surfaces S1 to S3 generated from the invalid region (blank portion) shown in the left figure of each figure. ..

In the present embodiment, the point cloud data is processed for each measurement area without extracting or generating the planar shape after all the measured point cloud data are coordinate-converted and then combined. That is, the point cloud processing step Step 3 is processed for each measurement step Step 2. Therefore, the amount of data processed once in the point cloud processing step Step 3 can be reduced, and the processing time for extracting or generating a planar shape can be shortened.

The three-dimensional model generation step Step 5 is, for example, a step of generating a three-dimensional model from the planar shapes shown in FIGS. 5A to 6C. Specifically, the process is performed based on the flow shown in FIG.

As shown in FIG. 4, the three-dimensional model generation step Step 5 includes, for example, the extraction surface synthesis step Step 51 in which the extraction surface is converted into the robot coordinate system and synthesized, and the assumption that the assumed surface is converted into the robot coordinate system and synthesized. The surface synthesis step Step 52, the plate thickness shape confirmation step Step 53 that processes a flat surface having a width corresponding to the plate thickness of the steel plate as the end face of the steel plate, and the shadow portion shape restoration that extracts and restores the shadow portion where the extraction surface and the assumed surface are not formed. The process Step 54 is provided.

The extraction surface synthesis step Step 51 is a step of synthesizing the extraction surface extracted by the point cloud processing step Step 3. Specifically, the extraction surface synthesis step Step 51 includes the first step Step 511 for acquiring all the extraction surface shapes, the second step Step 512 for converting the extraction surface shape into a robot coordinate system, and the extraction surface being the same as the registered surface. The third step Step 513 for confirming whether or not it is a surface area, the fourth step Step 514 for registering the extracted surface as a new surface when the extracted surface is not the same surface area as the registered surface, and the extracted surface have been registered. It is provided with a fifth step Step 515 for recalculating the plane / shape by superimposing the same surface region as the surface (Y), and a sixth step Step 516 for confirming whether or not all the extracted surfaces have been processed. ..

For example, since the planar shapes of the first measurement area to the ninth measurement area shown in FIGS. 5A to 6C are sensor coordinate systems, the viewpoint differs depending on each measurement area. Therefore, the coordinate system is unified by converting the coordinates of the sensor coordinate system into the coordinates of the robot coordinate system. Here, FIG. 7 is an image diagram showing the extraction surface synthesis step, (A) shows the synthesis method, and (B) shows the synthesis result.

The two planar shapes shown in the upper part of FIG. 7 (A) are the planar shapes of the first measurement region and the second measurement region shown in FIGS. 5 (A) and 5 (B). In the extraction surface synthesis step Step 51, a case where processing is performed in order from the first measurement region will be described. The plane shape included in the first measurement area A1 is coordinate-converted into the robot coordinate system, and all the planes included therein are registered in the motion automatic generation system as new planes. For example, as shown in FIG. 7A, the surfaces M1 to M4 are registered as new surfaces.

Next, the plane shape of the second measurement area A2 is coordinate-converted to the robot coordinate system, and it is confirmed whether or not the plane has the same surface area as the registered surfaces M1 to M4 registered in the motion automatic generation system. Now, as shown in FIG. 7 (A), if the plane included in the second measurement region A2 has the same plane region as the registered plane M1, the coordinates thereof are used in FIG. 7 (A). As shown in the lower part, the planar shape included in the first measurement area A1 and the planar shape included in the second measurement area A2 are combined. The surfaces that do not form the same surface area as the registered surfaces M1 to M4 are registered in the automatic motion generation system as new surfaces.

Hereinafter, by repeating this process for all the measurement regions, a pre-three-dimensional model in which all the extraction surfaces are synthesized can be generated as shown in FIG. 7 (B). Here, the case where the pre-three-dimensional model is generated from the plane-shaped extraction planes included in the first measurement region to the ninth measurement region shown in FIGS. 5 (A) to 6 (C) is illustrated.

The assumed surface synthesis step Step 52 is a step of synthesizing the assumed surface generated by the point cloud processing step Step 3. Specifically, the assumed surface synthesis step Step 52 confirms the effectiveness of the first step Step 521 that acquires all the assumed surface shapes, the second step Step 522 that converts the assumed surface shape into the robot coordinate system, and the assumed surface. It includes a third step Step 523, a fourth step Step 524 for synthesizing assumed surfaces, and a fifth step Step 525 for confirming whether or not all assumed surfaces have been processed.

The third step Step 523 is a step of confirming whether or not the assumed surface is recognized as a normal plane in another measurement area. For example, if there is an assumed surface in the visual field area created by the sensor origin of another measurement area and the outline of the extraction surface, the assumed surface is invalidated and excluded from the synthesis target. The fourth step Step 524 is processed by substantially the same process as the above-mentioned extraction surface synthesis step Step 51.

Here, FIG. 8 is an image diagram showing the assumed surface synthesis process. The assumed surface S1'shown in the upper part of FIG. 8 is an enlarged view of the assumed surface S1 included in the seventh measurement region after coordinate conversion. Further, the assumed surface S2'shown in the middle of FIG. 8 is obtained by converting the coordinates of the assumed surface S2 included in the eighth measurement region and synthesizing it with the assumed surface S1. Further, the assumed surface S3'shown in the lower part of FIG. 8 is obtained by converting the coordinates of the assumed surface S3 included in the ninth measurement region and synthesizing it with the assumed surface S2'.

The plate thickness shape confirmation step Step 53 is a step of performing a process of estimating a plane from the plate thickness shape. Since it is considered that the portion of the steel plate that can be regarded as the plate thickness shape forms the end face of the plane, the plane forming the end face can be estimated. For example, two vertical planes (a plane having an outer shape in the longitudinal direction close to parallel) that can be regarded as a plate thickness, up to a plane having a member width within the same plane normal direction and a possible structural range (a plane having an outer shape in the longitudinal direction). (Both sides of the plate thickness) are created from the longitudinal outline. By the process of the plate thickness shape confirmation step Step53, the floating surface on the three-dimensional model can be erased.

Here, FIG. 9 is an image diagram showing a plate thickness shape confirmation process. The figure shown in the upper part of FIG. 9 is a three-dimensional model after processing the extraction surface synthesis step Step 51 and the assumed surface synthesis step Step 52. The virtual surface L1 in the figure indicates an upper measurement boundary surface, and the virtual surfaces L2 and L3 indicate a side measurement boundary surface. The figure shown in the upper part of FIG. 9 is a three-dimensional model in which the plate thickness shape portion D is extracted from the upper part and the planar shape is estimated and added.

The shadow portion shape restoration step Step 54 is a step of restoring the shape of the portion that is in the shadow of the member and is not in contact with the surface. For example, it is restored by extending the shape of the intersecting side at an angle assumed in a nearby plane. By the process of the shadow portion shape restoration step Step 54, the unnaturally separated surfaces on the three-dimensional model can be erased.

Here, FIG. 10 is an image diagram showing a shadow portion shape restoration process. The figure shown in the upper part of FIG. 10 is a three-dimensional model after processing the extraction surface synthesis step Step 51 and the assumed surface synthesis step Step 52. In the figure, a portion where no surface is generated, that is, a shadow portion H is formed in the portion surrounded by the dotted line. The shadow portion H is restored as shown in the lower part of FIG. 10 by the above-mentioned processing of this portion.

The work setting step Step 6 is a step of inputting data necessary for welding such as a welded portion and a leg length. For example, in the case of welding a welding place having the same structure as an already welded place, since the data necessary for welding such as the welding place and the leg length are known, the work setting step Step 6 is the setting step Step 1. It may be processed before.

The welding motion generation step Step 7 is a step of generating a welding motion of the welding robot 1 (articulated arm 2) based on the three-dimensional model generated by the above-mentioned process. In the welding operation generation step Step 7, the inclination of the welding robot 1 is calculated using the data of the acceleration sensor, the inclination sensor, etc. arranged on the pedestal 3, and the operation is generated using the welding conditions of the welded joint in consideration of the direction of gravity. You may do so.

Further, the welding operation generation step Step 7 performs processing such as determining the posture of the welding robot 1, confirming the operation / interference of the articulated arm 2, moving to the welding start position, sensing operation, and welding operation for each gap. , Create operation data such as movement to the retracted position. Since the welding motion generation step Step 7 is substantially the same as the motion generation process of a general robot provided with an articulated arm, detailed description thereof will be omitted here.

In the construction order setting step Step8, in consideration of the drop of the welding slag, for example, the vertical welding is prioritized over the horizontal welding, or the lower welding is prioritized over the upper welding. This is the process of setting the welding order.

The welding work step Step 9 is a step of welding by the welding robot 1 based on the generated welding operation and construction order. For example, the welding robot 1 is placed at a predetermined position, the welding torch 2w is moved to the start position, the joint is sensed, and then welding is performed according to the gap length.

According to the above-described method for automatically generating the operation of the welding robot 1 according to the present embodiment, since the 3D measurement sensor 4 is used to generate a three-dimensional model of the welding location, the welding robot uses CAD data in advance. It is not necessary to generate the welding operation of 1. Therefore, according to the present embodiment, the preliminary work of the welding robot 1 can be reduced, and the operation of the welding robot 1 can be automatically generated at the site.

Further, the program for causing an arithmetic unit such as a computer to execute the operation automatic generation method of the welding robot 1 according to the present embodiment is a storage of SSD (Solid State Drive), HDD (Hard disk drive), etc. in the control box 5. It may be stored in the device, may be recorded on a recording medium readable by the reader arranged in the control box 5, or may be stored in the external control box side connected to the control box 5 by a cable. It may be executed by the arithmetic unit on the external control box side. Further, the program may be configured to be installable in the arithmetic unit via a network such as the Internet.

The recording medium is, for example, a USB (Universal Serial Bus) memory equipped with a semiconductor memory such as a flash memory. The recording medium may be a magnetic disk or an optical disk. Optical discs are, for example, CDs (Compact Discs) and DVDs (Digital Versailles Discs).

It goes without saying that the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention.

1 Welding robot 2 Articulated arm 2t Rotating table 2w Welding torch 3 Pedestal 4 3D measurement sensor 5 Control box 6 Handle 21 Base 22 Upper arm 23 Lower arm 24 Wrist 25 Tool part 31 Handle 32 Laser pointer 33 Leg part Step1 Setting process Step2 Measurement Process Step3 Point group processing process Step31 Data acquisition process Step32 Coordinate conversion process Step33 Local plane calculation process Step34 Labeling process Step35 Shape extraction process Step36 Invalid region extraction process Step37 Assumed surface generation process Step4 Measurement confirmation process Step5 Three-dimensional model generation process Step51 Extraction surface synthesis Step 511 1st step Step 512 2nd step Step 513 3rd step Step 514 4th step Step 515 5th step Step 516 6th step Step 52 Assumed surface synthesis step Step 521 1st step Step 522 2nd step Step 523 3rd step Step 524 4th step Step 525 Step53 Plate thickness shape confirmation process Step54 Shadow part shape restoration process Step6 Work setting process Step7 Welding operation generation process Step8 Construction order setting process Step9 Welding work process

According to the present invention, there is a method for automatically generating a welding motion of a welding robot placed at a predetermined welding place, which is a setting process for arranging a 3D measurement sensor at a predetermined place. The welding site is divided into a plurality of measurement areas and measured by the 3D measurement sensor, and an extraction surface is generated from the point group data that can be recognized as a plane among the measured point group data, and the point group data is generated. A point group processing step for generating an assumed surface from an unrecognizable invalid region that cannot be obtained, a three-dimensional model generation step for generating a three-dimensional model of the welding location from the extracted surface and the assumed surface, and the three-dimensional model. Provided is an automatic motion generation method for a welding robot, comprising a welding motion generation step of generating a welding motion of a welding robot using the welding robot , wherein the point group processing step is processed for each measurement step. Will be done.

Further, according to the present invention, it is a method of automatically generating a welding motion of a welding robot placed at a predetermined welding place, and is a setting step of arranging a 3D measurement sensor at a predetermined place. And, the measurement process of measuring the welding place with the 3D measurement sensor and the point group data that can be recognized as a plane among the measured point group data are generated to generate an extracted surface, and the point group data cannot be obtained and cannot be recognized. A point group processing step that generates an assumed surface from an invalid region, a three-dimensional model generation step that generates a three-dimensional model of the welding location from the extracted surface and the assumed surface, and welding of a welding robot using the three-dimensional model. The point group processing step includes a welding operation generation step for generating an motion, and the point group processing step recognizes that the invalid region is formed by a plane perpendicular to the extraction surface and generates the assumed surface. A method for automatically generating an operation of a welding robot is provided.

The three-dimensional model generation step includes an extraction surface synthesis step of converting the extraction surface into a robot coordinate system and synthesizing it, and a hypothetical surface synthesis step of converting the assumption surface into a robot coordinate system and synthesizing the surface. May be good.

The three-dimensional model generation step may include a plate thickness shape confirmation step of processing a flat surface having a width corresponding to the plate thickness of the steel plate as an end face of the steel plate.

The three-dimensional model generation step may include a shadow portion shape restoration step of extracting and restoring a shadow portion on which the extraction surface and the assumed surface are not formed.

The welding motion generation step may include generating the welding motion in consideration of the inclination of the place where the welding robot is arranged.

Further, according to the present invention, it is an motion automatic generation system of a welding robot that automatically generates a welding motion of a welding robot arranged at a predetermined welding place, and the welding place is divided into a plurality of measurement areas. A 3D measurement sensor for measurement and a calculation device for generating a three-dimensional model of the welding location based on the data of the 3D measurement sensor are provided, and the calculation device is measured for each measurement process of the 3D measurement sensor. An extracted surface is generated from the point group data that can be recognized as a plane among the point group data, an assumed surface is generated from an unrecognizable invalid region where the point group data cannot be obtained, and the welding is performed from the extracted surface and the assumed surface. Provided is an automatic motion generation system for a welding robot, which is configured to generate a three-dimensional model of a place and to generate a welding motion of the welding robot using the three-dimensional model.

Further, according to the present invention, it is a motion automatic generation system of a welding robot that automatically generates a welding motion of a welding robot placed at a predetermined welding place, and is a 3D measurement sensor for measuring the welding place. A calculation device that generates a three-dimensional model of the welding location based on the data of the 3D measurement sensor is provided, and the calculation device is an extraction surface from the point group data that can be recognized as a plane among the measured point group data. Is generated by recognizing that the assumed surface is formed by a plane perpendicular to the extraction surface from the unrecognizable invalid region where the point group data cannot be obtained, and the above is generated from the extraction surface and the assumption surface. Provided is an automatic motion generation system for a welding robot, which is configured to generate a three-dimensional model of a welding location and generate a welding motion of the welding robot using the three-dimensional model.

Claims (9)

It is a method of automatically generating the welding motion of a welding robot that automatically generates the welding motion of the welding robot placed at a predetermined welding place.
The setting process to place the 3D measurement sensor in place, and
A measurement process in which the welding location is divided into a plurality of measurement areas and measured by the 3D measurement sensor, and
A point cloud processing step that generates an extracted surface from the point cloud data that can be recognized as a plane from the measured point cloud data, and generates an assumed surface from an unrecognizable invalid area where the point cloud data cannot be obtained.
A three-dimensional model generation step of generating a three-dimensional model of the welding location from the extracted surface and the assumed surface, and
A welding motion generation process for generating a welding motion of a welding robot using the three-dimensional model, and a welding motion generation process.
A method for automatically generating motions of a welding robot, which comprises. The method for automatically generating an operation of a welding robot according to claim 1, wherein the point cloud processing step is processed for each measurement step. The method for automatically generating an operation of a welding robot according to claim 1, wherein the setting step is processed so that the plurality of measurement areas have an area overlapping with an adjacent measurement area. The method for automatically generating an operation of a welding robot according to claim 1, wherein the point cloud processing step recognizes that the invalid region is formed by a plane perpendicular to the extraction surface and generates the assumed surface. The three-dimensional model generation step includes a claim surface synthesis step of converting the extracted surface into a robot coordinate system and synthesizing it, and a hypothetical surface synthesis step of converting the assumed surface into a robot coordinate system and synthesizing the surface. Item 1. The method for automatically generating an operation of a welding robot according to Item 1. The method for automatically generating an operation of a welding robot according to claim 1, wherein the three-dimensional model generation step includes a plate thickness shape confirmation step of processing a flat surface having a width corresponding to the plate thickness of the steel plate as an end face of the steel plate. The method for automatically generating an operation of a welding robot according to claim 1, wherein the three-dimensional model generation step includes a shadow portion shape restoration step of extracting and restoring a shadow portion on which the extraction surface and the assumed surface are not formed. The method for automatically generating an operation of a welding robot according to claim 1, wherein the welding operation generating step includes generating the welding operation in consideration of an inclination of a place where the welding robot is arranged. It is a welding robot motion automatic generation system that automatically generates the welding motion of a welding robot placed at a predetermined welding location.
A 3D measurement sensor that divides the welding location into a plurality of measurement areas and measures them.
It is equipped with an arithmetic unit that generates a three-dimensional model of the welding location based on the data of the 3D measurement sensor.
The arithmetic unit generates an extraction surface from the point cloud data that can be recognized as a plane among the measured point cloud data, and generates an assumed surface from an unrecognizable invalid area where the point cloud data cannot be obtained, and the extraction surface. And, a three-dimensional model of the welding place is generated from the assumed surface, and the welding motion of the welding robot is generated by using the three-dimensional model.
An automatic motion generation system for welding robots.

Images