JP2017071033A - Working reference object, working reference object manufacturing method, robot arm adjusting method, vision system, robot apparatus, and indicator member - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide working reference objects capable of using positional deviation information of indicators even when a robot apparatus cannot acquire, from a server, indicator positional deviation information measured for each working reference object.SOLUTION: A tray 7 includes: a marker 11 which is optically detected to allow adjustment of a coordinate system used for control of a robot arm 2; and an error information pattern 12 which is optically detected to allow obtaining of positional deviation information of the marker 11 relative to the tray 7. An error information measurement process measures positional deviation information of an indicator in the tray 7. An error information recording process converts the indicator positional deviation information measured by the error information measurement process into the error information pattern 12, an example of optically readable information image, and then records the pattern in the tray 7. A robot apparatus 1a adjusts the coordinate system used for control of the robot arm 2, based on positional information and positional deviation information of the marker 11 obtained by imaging the marker 11 and the error information pattern 12 with a fixed camera 5.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a work reference object used in a robot apparatus having a robot arm, a work reference object manufacturing method, a robot arm adjustment method, a vision system, a robot apparatus, and an index member.

  A robot apparatus having a multi-degree-of-freedom robot arm is widely used. In robotic devices, trays that house multiple parts that are taken in and out by a robot arm with multiple degrees of freedom and jigs that temporarily hold the parts and work with the robot arm are used as work reference objects. The

  Here, when there are variations in accuracy between trays and jigs and tools, errors in putting in and out the parts by the robot arm and attachment work increase if the coordinate system for controlling the robot arm is kept constant. For this reason, in Patent Document 1, an index is provided on the work reference object, the image is captured by the camera, and the coordinate system for controlling the robot arm is corrected in accordance with the detected position of the index.

  However, in the case where the index itself is provided with a positional deviation with respect to the work reference object, an error occurs when the coordinate system for controlling the robot arm is adjusted in accordance with the misaligned index. Depending on the conditions, the error of the work by the robot arm on the work reference object may be increased.

  In order to cancel such positional deviation of the index itself, the positional deviation information of the index with respect to the work reference object is measured and corrected in advance for each work reference object. For example, Patent Document 2 discloses that a positional relationship between five marks is used in a measuring instrument in advance for a glass calibration jig for calibrating an assembling apparatus for laminating members.

  When there are a plurality of work reference objects, it is necessary to manage positional deviation information that differs for each individual in association with the actual object. The glass calibration jig described in Patent Document 2 is relatively easy to manage by hand because of the small number of individuals.

  However, for example, in the case where hundreds to thousands of trays are distributed in a factory, when managing individual difference information of trays that are work reference objects, it is possible to construct a management server. It is common. For example, the coordinate position information stored in the server is stored in the server in association with the code assigned to each work reference object, and the robot arm is controlled by calling the position displacement information from the server with the work reference object code to be used. The system will be adjusted.

Japanese Patent Laid-Open No. 7-84631 JP 2002-23851 A

  However, when the positional deviation information of the index with respect to the work reference object is called from the server to the robot apparatus, the robot apparatus that is not connected to the server cannot call the positional deviation information for each work reference object. In overseas and remote factories, it is not possible to call up positional information for each work reference object. When a large amount of work reference objects are replaced and used violently, the maintenance of the misregistration information management system becomes expensive.

  The present invention provides a working apparatus capable of adjusting a coordinate system for controlling a robot arm using the positional deviation information of an index even when the robot apparatus cannot acquire the positional deviation information of the index measured for each work reference object from the server. It is an object of the present invention to provide a reference object, a method for manufacturing a work reference object, a method for adjusting a robot arm, a vision system, a robot apparatus, and an index member.

  INDUSTRIAL APPLICABILITY The method for manufacturing a work reference object according to the present invention is used in a robot apparatus having a robot arm, and is an index in a work reference object having an index that can be optically detected and used to adjust the coordinate system used to control the robot arm. A measuring step of measuring the positional deviation information of the recording medium, and a recording step of converting the positional deviation information measured in the measuring step into an optically readable information image and recording it on the reference object for work. It is a manufacturing method.

  In the method for manufacturing a work reference object according to the present invention, the positional deviation information of the index for each work reference object is recorded as an information image on the manufactured work reference object. For this reason, even when the robot apparatus cannot acquire the positional deviation information from the server, it is possible to adjust the coordinate system that controls the robot arm by acquiring the positional deviation information of the index by reading the information image from the work reference object. A reference object can be provided.

FIG. 3 is an explanatory diagram of a configuration of the robot apparatus according to the first embodiment. It is explanatory drawing of a structure of a tray. It is explanatory drawing of a structure of the measurement recording device of error information. It is a flowchart of the error information recording process for the tray. It is an enlarged view around one marker of a tray. It is a flowchart of adjustment control of a robot arm. It is explanatory drawing of a robot operation correction calculation process. 10 is an explanatory diagram of a configuration of an error information measurement recording apparatus according to Embodiment 2. FIG. It is explanatory drawing of the structure of the robot apparatus of Embodiment 2. FIG. FIG. 10 is an explanatory diagram of a structure of a work holding unit in a second embodiment. It is a flowchart of teaching control of a robot arm. It is explanatory drawing of the two-dimensional code of marker position error data. It is explanatory drawing of the marker in the workpiece holder of Embodiment 3. FIG. It is explanatory drawing of the error information holding pattern in the workpiece holder of Embodiment 3. It is explanatory drawing of another example of a marker. It is explanatory drawing of another example of an error information pattern.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

<Embodiment 1>
(Robot device)
FIG. 1 is an explanatory diagram of the configuration of the robot apparatus according to the first embodiment. As illustrated in FIG. 1, in the robot apparatus 1 a, the control unit 4 controls the robot arm 2, the robot hand 3, and the fingers 31 to perform a picking operation on the workpiece 9 a in the tray 7. The robot apparatus 1a takes out the work 9a from the tray 7 by the robot arm 2 and transfers it to a subsequent process (not shown).

  The robot arm 2 is a 6-axis vertical articulated robot arm fixed to the gantry 10. A robot hand 3 that holds the workpiece 9 a is fixed to the tip of the robot arm 2. A finger 31 is fixed to the tip of the robot hand 3. The robot hand 3 grips and releases the workpiece 9 a by opening and closing the fingers 31.

  When the work 9a on the tray 7 is exhausted, the transport mechanism 401 ejects the tray 7 from which the work 9a has been removed, and transports a new tray 7 containing the work 9a into the movable range of the robot arm 2. The positioning of the new tray 7 is stopped in front of the robot apparatus 1a, and the picking operation is resumed by the robot arm 2, the robot hand 3, and the finger 31.

  The control unit 4 is a computer that controls the robot apparatus 1a. CPU41 performs a calculation. The ROM 42 and the RAM 43 store information. The interface unit (I / F) 44 communicates with the outside. The CPU 41, ROM 42, RAM 43, and interface unit (I / F) 44 can communicate with each other inside the control unit 4 via the bus 45.

  The CPU 41 functions as a robot control unit 411, a robot motion correction calculation unit 412, an image data acquisition unit (visual sensor control unit) 413, a marker position calculation unit 414, and an error information reading unit 415. The calculation unit and the control unit are temporarily formed in the CPU 41 by reading the program stored in the ROM 42 into the CPU 41. The program may be copied to the control unit 4 by a recording medium such as an optical disk.

  The robot arm 2, the robot hand 3, and the finger 31 are electrically connected to the robot control unit 411 of the CPU 41 via the interface unit 44. The robot control unit 411 outputs a command value, drives each joint of the robot arm 2, controls the position and orientation of the robot arm 2, and moves the tip of the robot arm 2 with six degrees of freedom. The robot control unit 411 adjusts the position and orientation of the robot hand 3 by outputting a command value. The robot control unit 411 outputs a command value to cause the finger 31 to open / close.

(Coordinate system)
A robot coordinate system R is set at the base end of the robot arm 2 as a coordinate system representing the position and orientation of the robot arm 2 with respect to the gantry 10. A hand coordinate system T representing the position and orientation of the robot hand 3 is set in the palm of the robot hand 3.

The robot control unit 411 calculates an angle target value to be taken by each joint of the robot arm 2 with respect to a target value (command value) of the relative position and orientation ^R H _T of the hand coordinate system T with respect to the robot coordinate system R. Do (inverse kinematics calculation). Here, the relative position and orientation from the robot coordinate system R to the hand coordinate system _T is represented as ^R H _T. In the present specification, the coordinate system symbol on the left shoulder and the lower right as the relative position and orientation and ^A H _B or ^A M _B coordinate transformation matrix representing the relative position and orientation of the coordinate system B from an arbitrary coordinate system A It expresses by giving.

The robot control unit 411 performs servo control so that the current angle output from an encoder (not shown) provided at each joint of the robot arm 2 matches the angle target value. The robot control unit 411 obtains the current angle information of each joint from the encoder to calculate the ^R H _T is the current relative position and orientation of the hand coordinate system T with respect to the robot coordinate system R (forward kinematics calculation).

  The robot apparatus 1a adjusts and visually adjusts the position of the robot arm 2 in order to automatically and accurately perform picking work and assembly work in the factory, or to adjust the position of teaching tools and workpieces with high accuracy and simplicity. Combined with sensor measurement.

  The robot apparatus 1a is equipped with a vision system that detects the tray 7 using a visual sensor. A visual sensor is a sensor that can measure the position and orientation of a measurement object two-dimensionally or three-dimensionally. Expressions such as “vision”, “robot vision”, and “camera” are used as words representing the visual sensor. In the first embodiment, the visual sensor is the fixed camera 5.

  As described above, the transport mechanism 401, which is an example of a holding unit, holds the tray 7 at a position where the robot arm 2 can perform processing. The fixed camera 5, which is an example of an imaging unit, can capture an area including at least the marker 11 and the error information pattern 12 on the tray 7 held by the transport mechanism 401. An error information reading unit 415, which is an example of an image processing unit, performs image processing on a captured image captured by the fixed camera 5, and acquires position information of the marker 11 and position shift information of the marker 11. The control unit 4, which is an example of an adjustment unit, adjusts the coordinate system used for controlling the robot arm 2 based on the acquired positional information of the marker 11 and positional deviation information of the marker 11.

(Fixed camera)
The control unit 4 performs image processing on the captured image of the fixed camera 5 and detects the state of the workpiece 9 a in the tray 7. The fixed camera 5 is installed at a position above the tray 7 and photographs the upper surface of the tray 7. The fixed camera 5 is electrically connected to the image data acquisition unit 413 of the CPU 41 via the interface unit 44.

  The image data acquisition unit 413 outputs an imaging command to the fixed camera 5 and the fixed camera 5 performs shooting. The image data acquisition unit 413 outputs a command to the fixed camera 5 to take an image of the tray 7, and transmits the image data from the fixed camera 5 to the image data acquisition unit 413. The control unit 4 acquires image data from the image data acquisition unit 413 and temporarily stores it in the RAM 43.

  Camera internal parameters and external parameters for converting pixel data of a captured image captured by the fixed camera 5 into three-dimensional coordinate data are acquired in advance and stored in the ROM 42. The image data acquisition unit 413 outputs the acquired image data of the captured image to the marker position calculation unit 414 and the error information reading unit 415.

  The marker position calculation unit 414 acquires the position information of the marker 11 by performing image processing on the image data of the captured image. The error information reading unit 415 performs image processing on the image data of the captured image, and acquires positional deviation information of the marker 11 with respect to the workpiece 9a in the tray 7.

  The robot motion correction calculation unit 412 adjusts the coordinate system for controlling the robot arm 2 using the position information and the position shift information of the marker 11. The robot control unit 411 operates the robot arm 2, the robot hand 3, and the finger 31 so as to cancel the positional deviation direction and the positional deviation amount of the marker 11 based on the calculation result of the robot motion correction calculation unit 412.

  A fixed camera coordinate system F representing the position and orientation of the fixed camera 5 is virtually set at the lens principal point position of the fixed camera 5. Calibration data for converting the three-dimensional coordinate data described from the fixed camera coordinate system F into the robot coordinate system R is also acquired in advance and stored in the ROM 42.

(tray)
FIG. 2 is an explanatory diagram of the configuration of the tray. 2A is a perspective view, and FIG. 2B is a plan view. As shown in FIG. 1, a robot apparatus 1a that is an example of a robot apparatus uses a tray 7 that is an example of a reference object for work. The tray 7, which is an example of a work reference object, is used in a robot apparatus having the robot arm 2.

  As shown in FIG. 2A, the tray 7 is a container that accommodates and transports the workpieces 9a in an aligned manner. The tray 7 is a container from which the workpiece 9a is taken out by the robot arm 2. The tray 7 can store 16 workpieces 9a by partitions. The tray 7 has a not-shown abutting portion with respect to the work 9a in each partition, and can position the work 9a with respect to the tray 7 within a predetermined accuracy range.

  As shown in FIG. 2B, the marker 11 and the error information pattern 12 are arranged adjacent to each other on the upper surface of the tray 7. The captured image of the fixed camera 5 includes four markers 11 on the upper surface of the tray 7 and eight error information patterns 12.

  As shown in FIG. 1, the tray 7 is preferably newly designed in accordance with the robot apparatus 1a. However, in the first embodiment, the tray 7 that has been used manually when the robot apparatus 1a is introduced, that is, the tray 7 that is not assumed to be measured by a visual sensor is used.

  Since it is not assumed to be measured by a visual sensor, it is difficult for the tray 7 to be measured by the fixed camera 5, and the marker 11 is likely to be required.

  In the tray 7, a plurality of markers 11, which are examples of indices, are arranged on one surface of the tray 7. The marker 11 can be optically detected and the coordinate system used for controlling the robot arm 2 can be adjusted. The marker 11 is realized by attaching a black circular seal to the tray 7.

(Error information recording process)
FIG. 3 is an explanatory diagram of the configuration of the error information measurement recording apparatus. FIG. 4 is a flowchart of error information recording processing for the tray. In the first embodiment, regardless of the operation of the robot apparatus 1a, error information is recorded on the tray 7 in advance separately using a dedicated error information measurement recording apparatus 300 shown in FIG. The control unit 304 includes a measurement control unit 313 and a recording control unit 314.

  As shown in FIG. 3, the measurement recording apparatus 300 uses the measurement microscope 305 and the marker 11 with respect to the reference position of the tray 7 in a state where the left side surface and the lower side surface of the tray 7 are mechanically abutted against the support portion 301. Measure the coordinate value of the center of the circle. In the error information measurement process which is an example of the measurement process, the measurement control unit 313 measures the positional deviation information of the index on the tray 7.

  In the measurement recording apparatus 300, the measurement control unit 313 performs image processing on the captured image of the tray 7 captured by the measurement microscope 305, acquires the positional deviation information of the marker 11, and outputs the positional deviation information as the error information pattern 12. 316 to output. In the error information recording process which is an example of the recording process, the error information pattern 12 which is an example of an information image in which the recording control unit 314 can optically read the index positional deviation information measured by the error information measurement process. And is recorded on the tray 7. Thereby, the tray 7 of Embodiment 1 is manufactured.

  As shown in FIG. 4 with reference to FIG. 3, the user prepares the tray 7 body before applying the marker 11, and sets it on the support portion 301 of the measurement recording apparatus 300 (S01). The user attaches the marker 11 to the tray 7 set on the support portion 301 (S02).

  The user sets the measurement dummy 309a of the workpiece (9a) on the tray 7 set on the support portion 301 and presses a measurement start button (not shown) of the measurement recording device 300. Then, the measurement recording apparatus 300 images the measurement dummy 309a and measures the amount of displacement of the attachment position of the marker 11 with respect to the reference position of the tray 7 (S03). The measurement dummy 309a is formed with an equidistant mesh for measurement and a center position. The measurement control unit 313 obtains a reference position where the marker 11 is to be arranged on the tray 7 from the captured image through the measurement microscope 305 of the tray 7 on which the measurement dummy 309a is set. Then, a positional deviation amount between the reference position and the current marker position is obtained for each of the four markers 11.

  The recording control unit 314 creates an error information pattern 12 that matches a definition described later as a label sticker, and prints it out from the output unit 316. The measurement recording device 300 calculates positional deviation information for each marker 11, creates an error information pattern 12 corresponding to the error amount (ΔX, ΔY) of the marker 11, and prints out the output from the output unit 316 (S04).

  The user pastes the error information pattern 12 for each marker 11 at a position close to the marker 11 of the tray 7 set on the support unit 301 (S05). In this way, the processing of the marker 11 and the error information pattern 12 for the tray 7 is completed.

  The position error of the marker 11 is an error that occurs for the first time by attaching the marker 11 to the tray 7. The position error information of the marker 11 is information generated for the first time when the marker 11 is attached to the tray 7. The error information pattern 12 reflects the position error information of the marker 11. The error information pattern 12 is recorded on the tray 7 according to the measurement result of the marker 11 after the marker 11 is provided on the tray 7.

  Therefore, the marker 11 and the error information pattern 12 are independent. “Independent” means that “the error information pattern 12 can be provided separately after the marker 11 has been processed”, “the positional deviation information held by the error information pattern 12 can be read separately from the position information of the marker 11. It means "I can do it."

  For example, the case where the marker 11 and the error information pattern 12 are simultaneously printed on one label sticker and attached to the tray 7 is not independent. However, the case where the error information pattern 12 is attached inside the marker 11 formed in a donut shape is independent.

(Error information pattern)
FIG. 5 is an enlarged view around one marker on the tray. In FIG. 5, (a) is a marker, (b) is an enlarged view of the marker, (c) is an error information pattern, and (d) is a detection range of the error information pattern.

As shown in FIG. 5A, the lower left point of the tray 7 is adopted as the reference position of the tray 7. As shown in FIG. 5B, the coordinate value of the measurement result viewed from the tray coordinate system W virtually set to the reference position of the tray 7 is (X _R , Y _R ), and the coordinate value design of the marker position is designed. Let the values be (X _D , Y _D ). At this time, the error amount (ΔX, ΔY) of the marker position is calculated as follows.

  As shown in FIG. 5C, the error information pattern 12 is defined as a rectangular shape having a length proportional to the error amount (ΔX, ΔY) of the marker position. The rectangular lengths ΔL1 and ΔL2 of the error information pattern 12 are defined to be 50 times the absolute value of the marker position errors ΔX and ΔY. Further, the attachment position of the error information pattern 12 with respect to the marker 11 corresponds to the error direction (that is, the sign of ΔX and ΔY).

  As shown in FIG. 2B, the tray 7 has an error information pattern 12 that is an example of an information image. The error information pattern 12 is disposed adjacent to the corresponding marker 11 on the surface of the tray 7 on which the marker 11 is disposed, and can be detected optically to acquire positional deviation information of the marker 11 with respect to the tray 7. The positional deviation information is the positional deviation direction and the positional deviation amount of the index with respect to the tray 7. The error information pattern 12 has a length corresponding to the coordinate axis direction component of the positional deviation amount in the direction of the coordinate axis corresponding to the positional deviation direction in the orthogonal coordinates centered on the marker 11.

(Coordinate adjustment to control the robot arm)
FIG. 6 is a flowchart of robot arm adjustment control. The control unit 4, which is a computer that has read the program, adjusts the work position on the tray 7 by the robot arm 2 using the error information pattern (12: FIG. 2).

  As shown in FIG. 6 with reference to FIG. 1, the control unit 4 detects that the tray 7 is transported within the movable range of the robot arm 2 by the transport mechanism 401 in the holding step (S11). In the holding step, the transport mechanism 401 holds the tray 7 at a position where the robot arm 2 can perform processing.

  In the detection step, the control unit 4 controls the fixed camera 5 with the image data acquisition unit 413 to photograph the tray 7 and the workpiece 9a (S12). The control unit 4 performs image processing on the image data acquired by the image data acquisition unit 413 by the marker position calculation unit 414, and acquires the position of the marker 11 (S13). In the detection step, the marker 11 of the tray 7 held by the transport mechanism 401 is detected by the fixed camera 5 which is an example of a detection unit, and the position of the index is acquired.

  In the reading process, the control unit 4 performs image processing on the image data acquired by the image data acquisition unit 413 by the error information reading unit 415 to acquire error information of the marker 11 (S14). In the reading process, the error information pattern 12 of the tray 7 held by the transport mechanism 401 is read by the fixed camera 5 which is an example of a reading unit, and the positional deviation information of the marker 11 is acquired.

  The control unit 4 uses a coordinate system for controlling the robot arm 2 using the position of the marker 11 acquired by the marker position calculation unit 414 and the error information of the marker 11 acquired by the error information reading unit 415 in the adjustment process. Is calculated (S15). In the adjustment step, the robot motion correction calculation unit 412 controls the robot arm 2 based on the position of the marker 11 acquired in the detection step and the positional deviation information of the marker 11 acquired in the reading step. Adjust.

  The control unit 4 controls the robot arm 2 and the robot hand 3 by the robot control unit 411 using the coordinate system corrected by the robot motion correction calculation unit 412 to perform the picking work of the work 9a in the tray 7 ( S16).

(Marker position calculation processing)
The marker position calculation unit 414 binarizes the entire image including the four markers 11 photographed by the fixed camera 5. Then, on the binarized image, the region of the marker 11 is specified from the shape feature amount such as roundness and circumscribed circle radius, and the edge image of the marker 11 is extracted. Thereafter, the center of the marker 11 is calculated with high accuracy by subjecting the edge image of the marker 11 to arc fitting processing.

  The marker position calculation unit 414 applies the internal parameters and external parameters of the fixed camera 5 that have been calibrated in advance to the coordinate data in units of pixels, thereby calculating the position of the marker 11 converted to an actual scale (for example, in mm). The value (X, Y) is calculated.

(Error information reading process)
The error information reading unit 415 performs image processing on the image data of the captured image of the tray 7 including the four error information patterns captured by the fixed camera 5, and measures the position and length of the pattern 12, thereby Error information of the marker 11 is acquired.

  As shown in FIG. 5D, the length of the error information pattern 12 can be measured by recognizing four areas centered on the marker 11.

  The presence / absence of the error information pattern 12 is detected by performing detection processing on the existence candidate region of the error information pattern 12 indicated by the dotted line with the position of the marker 11 acquired by the marker position calculation unit 414 as a reference. At this time, since the error information pattern 12 is detected on the −X side and the + Y side with respect to the marker 11, positive / negative information of the error is acquired thereby.

  Next, measurement is performed on the black rectangle inside the detected error information pattern 12, and the lengths L1 and L2 are calculated. Finally, the position error information of the marker 11 is calculated by performing inverse calculation of the length definition of the error information pattern 12.

  In this way, the error information reading unit 415 reads the position error information of the marker 11 embedded in the error information pattern 12 in one marker 11. The error information reading unit 415 detects the error information patterns 12 in the four markers 11 on the tray 7 shown in FIG. Read information.

(Robot motion compensation calculation processing)
FIG. 7 is an explanatory diagram of the robot motion correction calculation process. In FIG. 7, (a) is design data, (b) is measurement data, (c) is tray shape data, and (d) is correction result data.

  As shown in FIG. 1, the robot motion correction calculation unit 412 calculates the position and angle of the workpiece 9 a accommodated in the tray 7 by fitting processing at four points on the tray 7 corresponding to the four markers 11.

As shown in FIG. 7A, as a premise, the robot motion correction calculation unit 412 has design data. That is, the design value of the position of the marker 11 with respect to the tray coordinate system W _P D _{1 to P} D 4, a target position of the teaching point of the robot relative to the tray coordinate system W TP11～TP44, and tray coordinate system W to the robot coordinate system R Is the design position.

The marker position data calculated by the marker position calculation unit 414 is a measurement value seen from the coordinate system of the fixed camera 5 and cannot be used as it is for correction of the robot. Therefore, the robot motion correction calculation unit 412 performs coordinate conversion on the marker position data calculated by the marker position calculation unit 414, and the four marker 11 in the robot coordinate system R as shown in FIG. converted into the coordinate value _{P M 1~P} M 4.

  It should be noted that the relationship between the robot coordinate system R and the fixed camera coordinate system F specified in advance necessary for this purpose is acquired in advance by the calibration described above.

  The robot motion correction calculation unit 412 receives the position error information of the marker 11 output from the error information reading unit 415, and calculates marker position data in which the position error information is added to the design data.

Specifically, as shown in FIG. 7A, position error information (ΔX [i], ΔY [i]) (i = 1, 2, 3, 4) with respect to the four markers 11 is obtained. Correction is made by adding to the design values P _D 1 to P _D 4 of the design marker positions. The coordinate values of the corrected marker position data viewed from the tray coordinate system W are P _R 1 to P _R 4.

Robot operation correction calculation unit 412, the marker position data after correction viewed tray coordinate system W coordinate values _P R 1 to P of _R 4, the coordinate value _P M 1 to P _M of the marker measured data viewed from the robot coordinate system R The coordinate transformation parameters (tx, ty, θ) for coordinate transformation so as to be superimposed on 2 are obtained.

As specific processing, as shown in FIG. 7C, coordinate values after coordinate conversion of coordinate values P _R 1 to P _R 4 with coordinate conversion parameters (tx, ty, θ) and marker measurement are performed. as the sum of squares of distances of the coordinate values of the coordinate values _P M 1 to P _M 4 data is minimized, obtaining the coordinate conversion parameters (tx, ty, θ).

That is, when the coordinate transformation by the coordinate transformation parameters (tx, ty, θ) is expressed as a homogeneous transformation matrix H (tx, ty, θ), the coordinate transformation parameters (tx, ty, θ) as shown in the following equation: Is obtained by the method of least squares. In the following equation, coordinate values P _M [i] and P _R [i] are converted into two-dimensional homogeneous vectors T (X _M [i], Y _M [i], 1), T (X _R [i], Y _R [i], expressed in 1).

  The robot motion correction calculation unit 412 performs a correction calculation by performing coordinate conversion on the teaching point data TP11 to TP44 of the robot using the coordinate conversion parameters (tx, ty, θ), and ends the process.

(Comparative example)
In Comparative Example 1, as shown in Patent Document 1, the teaching position data of the robot arm 2 is corrected with reference to marks provided on an object such as a tray, jig, tool, or workpiece.

  In the first comparative example, the processing position of the robot arm 2 is corrected with respect to an index provided on an object such as a tray, jig, tool, or workpiece. However, in reality, since there is a marker position error due to a processing error or the like for each individual object, it is difficult to correct with high accuracy.

  In Comparative Example 2, as shown in Patent Document 2, the position adjustment of the tray, jigs, and workpieces is performed with high accuracy using error data obtained by measuring the positional relationship of the alignment marks using a measuring machine in advance. In order to perform correction with high accuracy, it is necessary to store measurement data for each individual measured in advance in the server in order to cancel the position error of the marker.

  However, if this method is simply applied, it is necessary to manage and link the measurement data for each individual object difference, so there are many cases such as general-purpose trays in the factory, or addition due to consumption If the number of devices is large, the management load and the possibility of mistakes increase.

  If a network is managed by a server and a network between devices is built or a database of error information is built in a device, the system needs to be scaled up.

  As another method, a memory tag may be provided for each tray, and error data measured in advance may be recorded for each tray and read to use the error data. In this case, however, a memory tag writing device and reading device are required.

(Effect of Embodiment 1)
In the first embodiment, the error is corrected by reading both the position measurement marker and the error information holding pattern independently given to the object by the same visual sensor. For this reason, it is possible to realize a highly accurate robot operation by canceling marker attachment and machining position errors on the target object, while having a simple system configuration that does not include a dedicated sensor or server that stores error information. it can.

  In the first embodiment, the correction data cancels the position error of the marker 11 with respect to the tray 7 in addition to the correction of the position and orientation of the tray 7. For this reason, it is possible to work with the fingers 31 approaching the workpiece 9a with high accuracy.

  In the first embodiment, the work 9a can be picked up with high accuracy with a simple configuration without providing a server or database for managing the individual difference information of the tray 7.

  In the first embodiment, the error amount and the length of the error information holding pattern 12 are associated with each other, and the error direction and the direction of the error information holding pattern 12 are associated with each other. In addition, the amount and direction of error can be understood intuitively.

  In the first embodiment, the robot apparatus 1a capable of canceling the individual difference with respect to the marker 11 on the tray 7 and performing high-precision work can be realized with a simple configuration.

<Embodiment 2>
FIG. 8 is an explanatory diagram of the configuration of the error information measurement recording apparatus according to the second embodiment. FIG. 9 is an explanatory diagram of a configuration of the robot apparatus according to the second embodiment. In the second embodiment, as shown in FIG. 8, the marker 11 and the two-dimensional code 13 are attached to the work holder 8 using the measurement recording device 300. Then, as shown in FIG. 9, the marker 11 and the two-dimensional code 13 provided on the work holder 8 are imaged by the hand camera 6 to teach the robot arm 2. Since other configurations and control contents are the same as those in the first embodiment, in FIG. 8 and FIG. 9, the same configurations as those in the first embodiment are denoted by the same reference numerals as those in FIG. 3 and FIG. Description is omitted.

(Measurement recording device)
As shown in FIG. 8, the user records error information on the work holder 8 to which the marker 11 is attached, using the dedicated error information measurement recording apparatus 300. The measurement recording apparatus 300 uses the measurement microscope 305 and the marker 11 with respect to the reference position of the workpiece holder 8 in a state where the left side surface and the lower side surface of the workpiece holder 8 to which the marker 11 is attached are abutted against the support portion 301. Measure the center coordinate value of.

  The measurement control unit 313 performs image processing on the captured image of the work holder 8 captured by the measurement microscope 305 to acquire positional deviation information of the marker 11. The recording control unit 314 converts the positional deviation information into the two-dimensional code 13 and prints it out from the output unit 316. The user attaches the two-dimensional code 13 to the work holder 8 to which the marker 11 is attached.

(Robot device)
As shown in FIG. 9, in the robot apparatus 1 b, the control unit 4 controls the robot arm 2, the robot hand 3, and the fingers 31 to perform assembly work on the workpiece held by the workpiece holder 8. The robot apparatus 1b attaches a workpiece to the workpiece holder 8 by the robot arm 2 and executes an operation of assembling other parts to the workpiece.

(Hand camera)
As shown in FIG. 9, the robot apparatus 1 b uses a hand camera 6 instead of the fixed camera 5. The hand camera 6 is a two-lens stereo camera attached to the robot hand 3 at the tip of the robot arm 2.

  The hand camera 6 is attached to the robot hand 3 only during teaching of the robot arm 2 and can be detached from the robot hand 3 during assembly work on an actual work (9b: FIG. 10B). The hand camera 6 is calibrated as a single camera in advance, and an arbitrary feature point on the image with reference to the hand camera coordinate system V virtually set at one lens principal point of the two-lens camera. 3D coordinates can be measured. The hand camera 6 is pre-calibrated with the robot arm 2 and the relationship between the hand coordinate system T and the hand camera coordinate system V is calculated in advance.

(Work holder)
FIG. 10 is an explanatory diagram of the structure of the work holder in the second embodiment. In FIG. 10, (a) is a work non-mounting state, (b) is a work mounting state.

  As shown in FIG. 9, a work holder 8 is installed in the movable range of the robot arm 2 in the robot apparatus 1b. A work holder 8 which is an example of a work reference object is a jig to which a work 9b is attached by a robot arm 2. The workpiece holder 8 is detachably attached to the work table 403 by an attachment / detachment unit 402.

  As shown in FIG. 10A, the work holder 8 has abutting portions 81 and 82 and clamp portions 83 and 84. A plurality of markers 11 as an example of an index are arranged on one surface of the work holder 8.

  The workpiece holder 8 is provided with one two-dimensional code 13 including positional deviation information of the three markers 11 at the center of the three markers 11. The two-dimensional code 13 that is an example of the information image is arranged away from the marker 11 on the surface on which the marker 11 is arranged, and includes positional deviation information of the plurality of markers 11. The two-dimensional code 13 is character information (including numbers) that is two-dimensionally encoded.

  As shown in FIG. 10B, the clamp portions 83 and 84 clamp the workpiece 9b by a linear motion mechanism (not shown). As shown in FIG. 9, the robot apparatus 1 b places the workpiece 9 b on the workpiece holder 8 as shown in FIG. 10B by the robot arm 2 and the robot hand 3. Thereafter, the robot apparatus 1b causes the clamp portions 83 and 84 to perform a clamping operation, and holds the workpiece 9b between the clamp portions 83 and 84 and the abutting portions 81 and 82.

  Thus, the workpiece 9b is positioned with a predetermined accuracy and reproducibility with respect to the workpiece holder 8 with the abutting portions 81 and 82 as a reference. After the workpiece 9b is held by the workpiece holder 8, the robot apparatus 1b uses the robot arm 2 and the robot hand 3 to perform an operation of assembling another component (not shown) to the workpiece 9b.

(Accuracy of workpiece holder)
As shown in FIG. 9, when exchanging the workpiece holder, the robot apparatus 1b removes the workpiece holder 8 and attaches another workpiece holder intended for another workpiece to the robot apparatus 1b. There are a plurality of robot systems corresponding to the robot apparatus 1b in the factory, and a plurality of workpiece holders having the same design as the workpiece holder 8 exist. In some cases, the work holder 8 is attached to a robot apparatus at another location and used in accordance with the production situation or maintenance situation in the factory.

  In order for the robot apparatus 1b to accurately assemble another part to the workpiece 9b, it is necessary to adjust the teaching point of the robot arm 2 with high accuracy before the assembling work. In particular, immediately after the workpiece holder 8 is replaced, teaching is important because there is a difference in individual processing of the workpiece holder 8. Therefore, the second embodiment is directed to the adjustment of the teaching point in the robot apparatus 1b. The work position taught on the work holder 8 by the robot arm 2 is corrected by the positional deviation information decoded from the two-dimensional code 13.

(Robot arm teaching)
FIG. 11 is a flowchart of teaching control of the robot arm. In the second embodiment, teaching points that can be photographed by the user are registered in advance, and the control unit 4 controls and moves the robot arm 2 by a program.

  As shown in FIG. 11, the control unit 4 controls the robot arm 2 by the robot control unit 411, and moves the robot arm 2 to a position where the hand camera 6 can photograph the workpiece holder 8 (S21).

  The control unit 4 operates the hand camera 6 with the image data acquisition unit 413 to capture the image of the workpiece holder 8 and acquire the image data of the workpiece holder 8 (S22). The control unit 4 performs image processing on the image data of the work holder 8 by the marker position calculation unit 414, and calculates the position of the marker 11 (S23).

  The control unit 4 performs image processing on the image data by the error information reading unit 415, and reads the position error information of the marker 11 (S24). In the control unit 4, the robot motion correction calculation unit 412 performs robot correction calculation (S25).

The control unit 4 adjusts the position of the robot hand 3 by operating the robot using a later-described homogeneous transformation matrix ^R M _T calculated by the robot motion correction calculation unit 412 as a command value (S26). The control unit 4 registers the posture of the robot as a teaching point with the position of the robot hand 3 adjusted and positioned (S27).

(Marker position calculation processing)
The image data acquisition unit 413 executes processing for acquiring image data of the work holder 8 having the marker 11 and the two-dimensional code 13 as in the first embodiment. However, since the hand camera 6 is a two-lens stereo camera, two images are acquired for each shooting, and information on the focal length direction can be acquired.

  The marker position calculation unit 414 executes the process of calculating the position of the marker 11 on the image data acquired by the image data acquisition unit 413 as in the first embodiment. However, since the hand camera 6 is a two-lens stereo camera, the position of the marker 11 can be calculated three-dimensionally.

The marker position calculation unit 414 calculates a three-dimensional coordinate value for each of the three markers 11 using the calibration data of the hand camera 6 calibrated in advance. The homogeneous vector of the coordinate value is expressed by the following equation. In the following equation, the coordinate value P _M [i] is the coordinate value of the marker 11 as viewed from the hand camera coordinate system V.

(Error information reading process)
FIG. 12 is an explanatory diagram of a two-dimensional code of marker position error data. As illustrated in FIG. 9, the error information reading unit 415 performs image processing on the image data of the two-dimensional code 13 and acquires position error information of the marker 11. The error information reading unit 415 acquires error information by decoding the two-dimensional code 13 in the image data obtained by imaging the work holder 8.

As shown in FIG. 12, the two-dimensional code 13 stores marker position error data obtained by measuring the position error of the marker 11 measured in advance using the measurement recording apparatus 300 in the same manner as in the first embodiment. Yes. The marker position error data includes the design values P _D 1 to P _D 3 of the coordinate values of the three markers 11 and the coordinate values P _R 1 to P of the measurement values obtained by measuring the positions of the markers 11 with the measurement recording device 300 in advance. Defined as the difference of _R3 .

  When the two-dimensional code 13 is decoded, table data of error information for each point as shown in Table 1, for example, can be read.

By reading the two-dimensional code 13, error information ΔP [i] (i = 1, 2, 3) of the three markers 11 is acquired. Here, ΔP [i] is a 4 × 1 homogeneous vector having three-dimensional data. The value of the coordinate value _P R 1 to P _R 3 design value _P D 1 to P _D 3 and measurements, the abutting portions 81 and 82 of the workpiece retainer 8, workpiece mounting surface of the workpiece retainer 8 May be expressed as coordinate values viewed from the workpiece holder coordinate system J defined by

(Robot motion compensation calculation processing)
The robot motion correction calculation unit 412 uses the coordinate value P _M [i] (i = 1, 2, 3) of the marker measurement value output from the marker position calculation unit 414 as a specific method of the correction calculation. Coordinates are converted into coordinate values viewed from T. For this purpose, the pre-acquired relationship of the hand coordinate system T and the hand camera coordinate system V by the calibration is calculated as the following equation using the ^T H _V.

Here, the homogeneous transformation matrix from the robot coordinate system R to the hand coordinate system T is denoted as ^R H _T, and the homogeneous transformation matrix from the hand coordinate system T to the hand camera coordinate system V is denoted as ^T H _V. The coordinate system P _M [i] is described on the left shoulder of the character so that the coordinate value seen from which coordinate system is clear. ^V P _M [i] is the coordinate value of the marker measurement value viewed from the hand camera coordinate system V (that is, the coordinate value P _M [i] calculated in step S23), and ^R P _M [i] is the same point as the robot. This is a coordinate value viewed from the coordinate system R.

The homogeneous transformation matrix ^R H _T from the robot coordinate system R to the hand coordinate system T represents the robot posture at the time of shooting, and the robot controller 411 performs forward kinematics calculation from each joint angle of the robot arm 2. Etc.

Next, the robot motion correction calculation unit 412 corrects the marker design value information using the marker position error information. Design values P _D 1 to P _D 3 of the coordinate values of the marker 11 with respect to the workpiece holder coordinate system J are stored in the ROM 42 in advance. By adding the designed value _P D 1 to P error information ΔP error information reading unit 415 has read to _D 3 of the coordinate value [i] (i = 1,2,3) , the workpiece retainer currently used The coordinate values P _R 1 to P _R 3, which are 8 unique dimension values, are calculated. Since the coordinate values P _R 1 to P R ₃ are those described in reference to the workpiece retainer coordinate system J, can calculate the homogeneous vector as the following formula when you write ^{J P} R _[i].

Finally, by using the coordinate values ^J P _M based on a measured value of the marker 11 by the marker position calculating unit 414, a coordinate value ^J P _R in consideration of the position error of the marker 11 to the design value of the workpiece retainer 8, The position and orientation to be taken by the robot arm 2 are calculated.

Here, the position and orientation of the hand coordinate system T to be taken with respect to the workpiece holder coordinate system J is an item to be designed in advance, and its known coordinate transformation matrix is ^T M _J. In addition, in order to bring the relationship between the workpiece holder coordinate system J and the hand coordinate system T closer to ^T M _J , a three-dimensional homogeneous transformation matrix ^R M _T representing an unknown posture that the robot arm 2 should take is written. At this time, the homogeneous transformation matrix ^R M _T can be obtained by solving the following least squares method.

Unknown variables in the above formula is only homogeneous transformation matrix ^R M _T, the other variables are known by the processing so far. If there are three or more data not in the same straight line, it is possible to calculate the least-squares solution of homogeneous transformation matrix ^R M _T which is a six degrees of freedom unknown.

(Effect of Embodiment 2)
In the second embodiment, since the robot arm 2 can be taught by canceling the positional deviation error of the marker 11 with respect to the workpiece holder 8, the robot arm is more accurate than when the positional deviation error of the marker 11 is not canceled. 2 positioning is possible.

  In the second embodiment, since the two-dimensional code 13 provided at a position away from the marker 11 has the position error information of the three markers 11 collectively, the work individual difference information is stored in a server or database. This can be realized with a simple configuration that does not need to be provided.

<Embodiment 3>
FIG. 13 is an explanatory diagram of markers in the work holder of the third embodiment. FIG. 14 is an explanatory diagram of an error information holding pattern in the work holder of the third embodiment. In FIG. 13, (a) is a perspective view of the workpiece holder, (b) is a sectional view in the depth direction, and (c) is a perspective view of the marker pin. In FIG. 14, (a) shows an unrecorded state of error information, and (b) shows a recorded state of error information.

  In the second embodiment, the amount of positional deviation of the marker in the plane along the reference plane of the workpiece holder is two-dimensionally encoded and attached to the workpiece holder. In contrast, in the third embodiment, the amount of marker displacement in the height direction of the reference surface of the work holder is recorded on the marker as an error information holding pattern. Using the error information of the workpiece holder (8b: FIG. 13), the coordinate system in the depth direction of the marker (15: FIG. 13) is corrected. Since other configurations and control contents are the same as those in the second embodiment, description will be made with reference to FIGS.

(Work holder)
As shown in FIG. 13A, the work holder 8b, which is an example of a work reference object, has marker pins 15 inserted into the four reference holes 14, respectively. The user processes the reference hole 14 on the workpiece holder 8b with high accuracy, and then manually inserts the marker pin 15 manufactured with high accuracy into the reference hole 14. Regarding the fixing of the marker pin 15, the tolerance relationship between the reference hole 14 and the marker pin 15 may be designed so as to be tightly fitted and press-fitted. Alternatively, both may be loosely fitted and fixed with an adhesive or the like.

  The marker pin 15 which is an example of the index member is used in the robot apparatus 1a having the robot arm 2 and the work holder 8b. The fitting part 19 which is an example of a fitting part fits into the hole which the workpiece holder 8b has.

  The marker unit 16, which is an example of an index unit, can adjust the coordinate system used for controlling the robot arm 2 by optical detection. On the other hand, the error information entry unit 17, which is an example of the information image recording unit, can record an information image that can be optically read to acquire the positional deviation information of the index with respect to the work holder 8 b independently of the marker unit 16. .

  As shown in FIG. 13 (b), a marker unit 16 which is an example of an index that can be retreated from the reference surface of the workpiece holder 8b and can acquire position information in the retreat direction is disposed. The reference hole 14 has a stepped counterbore shape. The marker pin 15 is inserted into the reference hole 14 until the stepped portion of the marker pin 15 abuts in the depth direction. The marker pin 15 is fitted and positioned with a predetermined accuracy by the fitting portion 19 with respect to the reference hole 14.

  As shown in FIG. 13B, the reference hole 14 is designed such that the counterbore depth La is longer than the thickness Lb of the stepped portion of the marker pin 15. For this reason, the upper surface of the marker pin 15 does not protrude from the reference surface of the workpiece holder 8b, but rather recedes.

  As shown in FIG. 13C, the marker pin 15 is a cylindrical pin having a stepped shape. On the upper surface of the marker pin 15, a marker portion 16 having a cross mark shape and an error information entry portion 17 are formed. In the marker unit 16, the intersection of the cross marks coincides with the center axis of the marker pin 15 with a predetermined accuracy.

  As shown in FIG. 14A, the marker pin 15 has an error information entry part 17 arranged on the upper surface thereof so as to surround the marker part 16. The error information entry unit 17 has squares delimited by solid or dotted scales at regular intervals. In the error information entry section 17, one square represents a depth dimension of 10 μm, and a depth of 0 to 160 μm can be expressed by an area of 16 squares.

  In the third embodiment, the positional deviation information is the amount of retreat from the reference plane to the marker unit 16. The error information holding pattern 18, which is an example of an information image, is an arc having a length corresponding to the amount of retraction from the reference plane arranged surrounding the marker portion 16 to the marker portion 16.

(Marker error in XY direction)
As shown in FIG. 13A, the hole position of the reference hole 14 is processed with high accuracy based on a predetermined tolerance in the XY direction along the reference surface with respect to the mechanical reference of the workpiece holder 8b. Has been. When manufacturing the workpiece holder 8b, an appropriate machine tool and blade are used, and the four reference holes 14 are continuously machined using the same blade and machine tool, so that the XY directions along the reference surface The hole position is processed with high accuracy.

  Further, by appropriately managing the fitting tolerance between the reference hole 14 and the marker pin 15 in the fitting portion 19, the marker pin 15 is positioned with high accuracy with respect to the reference hole 14. In this way, the positional accuracy of the marker portion 16 with respect to the workpiece holder 8b in the XY directions along the reference plane achieves a high accuracy of about ± 10 μm.

(Marker error in the Z direction)
As shown in FIG. 13B, the error in the depth direction (Z direction) of the reference hole 14 of the marker portion 16 is larger by one digit than the error in the XY direction along the reference plane of the marker portion 16. Machining the counterbore depth La of the reference hole 14 with high accuracy for reasons such as tool change, compared to machining the hole position in the XY direction along the reference surface with high accuracy. Because it is difficult. This is because it is difficult to process the thickness dimension Lb of the stepped portion of the marker pin 15 with high accuracy due to the occurrence of tool replacement. This is because the reference hole 14 and the abutment surface in the depth direction of the marker pin 15 do not always abut against each other due to the presence of dust or the like.

  However, if the counterbore depth La of the reference hole 14, the thickness dimension Lb of the stepped portion, and the abutment in the depth direction of the reference hole 14 and the marker pin 15 are strictly specified, the manufacturing cost of the work holder 8 b increases. Therefore, the position accuracy of the marker portion 16 in the depth direction (Z direction) is designed to be low.

(Measurement recording device)
As shown in FIG. 8, the user records error information on a work holder (8b: FIG. 13) to which a marker (15: FIG. 13) is attached, using a dedicated error information measurement recording apparatus 300. The measurement recording device 300 measures the surface height of the marker (15: FIG. 13) provided on the workpiece holder (8b: FIG. 13) and the level difference between the reference surface of the workpiece holder (8b: FIG. 13).

  The measurement recording apparatus 300 acquires the step amount between the marker (15: FIG. 13) and the reference surface of the work holder (8b: FIG. 13), and displays the step amount on the output screen of the output unit 316. The user records the step amount displayed on the output screen in the error information entry section (17: FIG. 13) provided in the marker (15: FIG. 13).

  The user measures a depth dimension (a step amount) from the reference surface of the work holder 8b to the upper surface of the marker pin 15 (the surface on which the marker portion 16 is disposed). The level difference can be measured with high accuracy using a relatively simple measuring device. For example, the depth dimension may be measured by scanning a dial gauge from the upper surface of the work holder 8b to the upper surface of the marker pin 15 with reference to a precision surface plate (not shown). Moreover, you may measure using depth measuring instruments, such as a depth gauge.

  After measuring the depth of the upper surface of the marker pin 15 with respect to the reference surface of the workpiece holder 8b, the user makes a number of squares corresponding to the depth measurement value to the error information entry portion 17 of the marker pin 15. Fill with oily marker. For example, when the measured value of the depth of the upper surface of the marker pin 15 is 70 μm, the user fills the seven squares with the oil marker as shown in FIG.

  The user performs attachment of the marker pin 15, measurement of the height of the marker portion 16, and recording of error information in the error information entry portion 17 for each of the four reference holes 14 and the marker pins 15. The user processes the error information holding pattern 18 by recording the measurement result in the depth direction of the marker portion 16 in the error information entry portion 17 at the four reference holes 14. In this way, the processing of the error information holding pattern 18 is completed.

(Robot device)
In the third embodiment, as shown in FIG. 9, the robot apparatus 1 b is configured so that the control unit 4 controls the robot arm 2, the robot hand 3, and the fingers 31 to assemble the workpiece held by the workpiece holder 8. I do. The robot apparatus 1a performs the operation | work which attaches a workpiece | work to the workpiece holder 8 with the robot arm 2, and assembles other components with respect to a workpiece | work. Similar to the second embodiment, the region including the four marker portions 16 of the work holder 8 is imaged by the hand camera to acquire positional deviation information.

  When the coordinate system used for controlling the robot arm 2 is adjusted using the workpiece holder 8b, the four error information entry sections 17 of the workpiece holder 8 are imaged by the hand camera 6. The error information reading unit 415 performs image processing on the captured image of the work holder 8 and reads how many squares of the squares of the error information entry unit 17 are filled as the error information holding pattern 18.

  The robot motion correction calculation unit 412 corrects the coordinate system for controlling the robot arm 2 only in the Z direction with respect to the reference plane at the positions of the four marker pins 15.

  Note that reading error information is the same as in the first embodiment, and thus detailed description thereof is omitted. The correction of the coordinate system for controlling the robot arm 2 is the same as that in the second embodiment where the XY errors are 0, and thus detailed description thereof is omitted.

(Effect of Embodiment 3)
In the third embodiment, since the error measurement is completed only by the step measurement as the correction only in the Z direction, it can be performed using a simple measuring instrument such as a dial gauge or a depth gauge.

  In the third embodiment, the result measured with an oil marker or the like in the error information entry part is used as the error information holding pattern. For this reason, a special processing machine and a measuring machine are not required, and an assembly operator of a production apparatus can easily embed error information on site.

  In the third embodiment, when the robot arm teaching or the like is performed with six degrees of freedom using a three-dimensional measurement result by a stereo camera or the like, the error in the depth direction of each marker is the inclination of the hand coordinate system T of the robot arm. It becomes an error. The error in the tilt of the hand coordinate system T of the robot arm increases as the distance between the work holder 8b and the hand increases. For this reason, by improving the error in the Z direction of the marker portion 16, the positional accuracy in the XY directions of the hand coordinate system T of the robot arm with respect to the workpiece holder 8b is also improved.

  In the third embodiment, the positions of the marker portion 16, the error information entry portion 17, and the error information holding pattern 18 in the Z direction are lower by one step than the upper surface of the workpiece holder 8b. For this reason, there is a low risk that the marker unit 16 and the error information holding pattern 18 may be worn or damaged due to direct contact with the workpiece during production. Thereby, the error can be corrected over a long period of time with high reliability.

  In the third embodiment, since the amount of error is associated with the number of squares to be filled, the error state can be intuitively understood when the operator looks at the workpiece holder 8b.

<Other embodiments>
Embodiments 1, 2, and 3 show examples of the embodiment of the present invention, and can be appropriately changed without departing from the gist of the present invention. The present invention can be implemented in another embodiment in which part or all of the configurations of the first, second, and third embodiments are replaced with equivalent configurations.

(Marker shape)
FIG. 15 is an explanatory diagram of another example of the marker. In the first embodiment, the marker 11 is realized by attaching a black circular seal to the tray 7. In the second embodiment, the shape of the marker 11 is circular. In the third embodiment, the marker portion 16 is a cross mark surrounded by a circle. However, the marker may have any other shape such as a triangle or a rectangle.

  The marker 11 and the marker portion 16 may directly form a pattern on the metal material of the tray 7 and the work holder 8 by laser marking, printing, sandblasting, or the like. As long as the marker 11 and the marker portion 16 can be identified from the background tray 7 and the work holder 8 by the fixed camera 5, any means can be used. The marker 11 and the marker unit 16 may be formed by performing simple hole processing on the flat positions of the tray 7 and the workpiece holder 8.

  As shown in FIG. 15A, a cross mark surrounded by a circle may be adopted. As shown in FIG. 15B, a cross mark surrounded by a square may be used. As shown in (c) of FIG. 15, a square in which the diagonal direction is inclined at 45 ° may be used. The cross mark can easily obtain the coordinates of the center position by image processing.

(Error information pattern)
FIG. 16 is an explanatory diagram of another example of the error information pattern. In the first embodiment, the error information pattern 12 has a rectangular shape, and the lengths L1 and L2 of the rectangular shape are proportional to the marker position error amount. However, the configuration that the error information pattern 12 can take is not limited to this. It is sufficient that the error information reading unit 415 can read the position error information of the marker 11 from the image data of the error information pattern 12.

  The relationship between the error amount and the dimension of the error information pattern is not limited to a proportional relationship, and an arbitrary function having a correlation can be used so that the dimension of the error information pattern can be converted into the error amount.

  In the first embodiment, whether the error information pattern 12 is in the candidate area is determined whether the position error of the marker 11 is positive or negative. However, the size of the error information pattern 12 is simply changed from minus to plus. You may make it respond | correspond to error amount.

  As shown in FIG. 16A, the diameter of a circle can be used as position error information. As shown in FIG. 16B, the distance between the feature points 23a and 23b can be used as position error information. As shown in FIG. 16C, the distance between the sides 23e and 23f can be used as position error information.

  As shown in FIG. 16D, three or more feature points 23a, 23b, and 23c are provided in the error information pattern 24, and the midpoint 24d of the feature points 23a and 23b is set as a zero point. It is good also as a structure which judges plus or minus from a relative relationship. Arbitrary dimensions of characteristic figures other than these can also be used as position error information.

  In the second embodiment, the example in which the two-dimensional code 13 is used as the information image has been described. However, the error information of the marker position may be recorded by other methods. For example, a bar code can be used. It is also possible to print the numerical information as characters as they are on the flat surface of the tray 7 or the work holder 8 and read the character information from the captured image data using a known OCR technique.

  In the third embodiment, it is possible to arbitrarily set the number and shape of the squares of the error information entry unit 17 and the error amount represented per square. Further, instead of making the error amount proportional to the number of squares in the error information entry unit 17 that fills the squares, each square may be considered as bit information so that an error can be entered in binary. In this case, it is difficult for an operator to intuitively understand the amount of error, but there is an advantage that an error range that can be expressed in the same space is widened.

  In the first embodiment, the error information pattern 12 is recorded by printing. However, if the error information pattern 12 can be discriminated from the background of the tray 7 in the captured image of the fixed camera 5, another recording method such as laser marking, etching, sand blasting, drilling, or the like can be used.

(Reference object for work)
In the first embodiment, the tray 7 and the second and third embodiments have been described using the tray 7 and the work holder 8 as work reference objects. However, the work reference object may be any object that can be worked by the robot arm 2. It can be extended to things. The reference object for work may be a case, a tray, a through box, a standing board, a jig, a chuck, a tool, a measuring instrument, an assembly part, or the like.

  For example, the marker 11 may be provided on the body part itself to be assembled by the robot arm 2, the marker 11 of the body part may be measured, and the position error information of the marker 11 may be printed on the body part. Thereafter, the main body part may be imaged using the fixed camera 5 shown in FIG. 1, the position error information of the marker 11 may be read, and the attachment work of the attachment part to the main body part may be performed with high accuracy by the robot apparatus 1a.

  In the first, second, and third embodiments, the configuration in which the workpiece 9 a is gripped by opening and closing the fingers 31 is illustrated. However, it can be expanded to an arbitrary end effector attached to the robot arm 2 and acting on the workpiece 9a. For example, a suction pad that acquires the workpiece 9a by vacuum suction, an electromagnet tool that acquires the workpiece 9a using magnetic force, or the like may be used instead of the robot hand 3.

(Control part)
The control unit 4 shown in FIG. 1 may not be temporarily formed by reading the program stored in the ROM 42 into the CPU 41. The robot control unit 411, the robot motion correction calculation unit 412, the image data acquisition unit 413, the marker position calculation unit 414, and the error information reading unit 415 are limited to functions temporarily formed by a program as a computer as the control unit 4. Not.

  The control unit 4 may include the robot control unit 411, the robot motion correction calculation unit 412, the image data acquisition unit 413, the marker position calculation unit 414, and the error information reading unit 415 as individual circuits. You may implement | achieve the function of the equivalent control part 4 by combining the several control part provided separately.

  In the second embodiment, teaching points that can be photographed by the user are registered in advance, and the control unit 4 controls and moves the robot arm 2 by a program. However, the robot arm 2 may be moved by inching using a teaching pendant (not shown).

  In the second embodiment, the case where the correction operation is executed only once has been described, but the error is converged by repeating the jig imaging S22 to the position / orientation adjustment operation S26 a plurality of times for the same workpiece holder 8. You may adopt the method of letting you do. In this case, since the position error information of the marker 11 does not change, the error information reading (S24) by the error information reading unit 415 can be omitted after the second time. Finally, after the position / orientation adjustment operation S26 is completed, imaging may be performed again, and the teaching point storage process (S27) may be executed after confirming that positioning is performed within a predetermined error range.

(Error measurement)
As a method for calculating the position of the marker 11 with image processing, various methods are known in addition to arc fitting processing of the edge image. For this reason, any known method may be used as a method of calculating the position of the marker 11.

  For the measurement of the error amount (ΔX, ΔY) of the marker position of the marker 11, it is possible to use a known method other than image-wise measurement using a measurement microscope. As another known method, with respect to the marker printed on the surface of the tray 7, the tray 7 set on the support unit 301 is imaged with a fixed camera, and the captured image is image-processed to perform an error amount (ΔX , ΔY) can also be measured. As another known method, when a machine shape such as a hole is used as the marker 11, it can be measured by a contact type three-dimensional measuring instrument.

  Specific examples of visual sensors include general monocular industrial cameras, stereo cameras with multiple cameras for three-dimensional measurement, optical systems that project sheet-like laser light and pattern light onto objects, 3D measurement sensors combined with cameras are included.

In the first embodiment, the case where the displacement amount of the attachment position of the marker 11 is measured using the measurement dummy 309a has been described. However, in the measurement of the positional deviation amount, the measurement value (X _R , Y _R ) of the position of the marker 11 is obtained with reference to the contact position of the tray 7 with respect to the support portion 301 without using the measurement dummy 309a. The amount of deviation may be calculated.

  In the first embodiment, the example in which the control unit 4 detects the marker 11 and the error information pattern 12 and adjusts the coordinates for controlling the robot arm 2 has been described. However, this is substantially the same as the embodiment in which the control unit 4 detects the marker 11 and the error information pattern 12 and adjusts the target position of the work on the tray 7 by the robot arm 2.

1a, 1b Robotic device (robot system)
2 Robot arm, 3 Robot hand, 4 Control unit 5 Fixed camera, 6 Hand camera, 7 Tray 8 Work holder, 9a, 9b Work 10 Mount, 11 Marker, 12 Error information pattern 13 Two-dimensional code, 14 Reference hole 15 Marker Pin, 16 Marker part, 17 Error information entry part 18 Error information holding pattern, 19 Fitting part 31 Finger, 41 CPU, 42 ROM, 43 RAM
44 interface unit, 45 bus, 81, 82 abutting unit 83, 84 clamping unit 300 measurement recording device, 301 support unit, 304 control unit 305 measurement microscope, 401 transport mechanism, 402 detachable unit 411 robot control unit, 412 robot operation correction Calculation unit 413 Image data acquisition unit, 414 Marker position calculation unit 415 Error information reading unit

Claims (17)

A measuring step used in a robot apparatus having a robot arm to measure positional deviation information of an index in a working reference object having an index that can be optically detected and adjusts a coordinate system used for controlling the robot arm;
And a recording step of converting the positional deviation information measured in the measuring step into an optically readable information image and recording the information image on the working reference object. . A reference object for work used in a robot apparatus having a robot arm,
An index capable of optically detecting and adjusting a coordinate system used for controlling the robot arm;
A work reference object, comprising: an information image that can be optically read to obtain positional displacement information of an index with respect to the work reference object. A plurality of the indicators are arranged on one surface of the working reference object,
The work reference object according to claim 2, wherein the information image is arranged adjacent to the corresponding index on the one surface. The positional deviation information is a positional deviation direction and a positional deviation amount of the index with respect to the work reference object,
4. The information image according to claim 3, wherein the information image has a length corresponding to a coordinate axis direction component of the positional deviation amount in a direction of a coordinate axis corresponding to the positional deviation direction in orthogonal coordinates centered on the index. The reference object for work described. A plurality of the indicators are arranged on one surface of the working reference object,
3. The work reference object according to claim 2, wherein the information image is arranged apart from the index on the one surface and includes a plurality of misalignment information of the index.   The work reference object according to claim 5, wherein the information image is two-dimensionally encoded numerical information. An indicator that can be retreated from the reference surface of the working reference object to obtain position information in the backward direction is arranged,
The positional deviation information is a retraction amount from the reference plane to the index,
The work reference object according to claim 2, wherein the information image is formed on the same member as the index and is disposed at a position retracted from the reference surface of the work reference object.   The work reference object according to any one of claims 2 to 7, wherein the work reference object is a container from which a workpiece is taken out by the robot arm.   The work reference object according to any one of claims 2 to 7, wherein the work reference tool is a jig to which a workpiece is attached by the robot arm. 10. A working reference object holding unit, a detection unit for optically detecting an index, a reading unit for optically reading an information image, and a robot arm control unit, and a robot arm control unit. A method for adjusting a robot arm in a robot apparatus that uses the work reference object according to claim 1,
A holding step in which the control unit holds the reference object for work at a position where the processing by the robot arm can be performed by the holding unit;
A detecting step in which the control unit detects the index of the reference object for work held in the holding step, and acquires the position of the index by the detecting unit;
A reading step in which the control unit reads the information image of the reference object for work held in the holding step by the reading unit to acquire positional deviation information of the index;
The control unit includes an adjustment step of adjusting a coordinate system used for controlling the robot arm based on the position of the index acquired by the detection step and the positional deviation information of the index acquired by the reading step. A method for adjusting a robot arm as a feature.   A program for causing a computer to execute the robot arm adjustment method according to claim 10.   The recording medium which recorded the program of Claim 11. A vision system for detecting the work reference object according to any one of claims 2 to 9,
An imaging unit capable of imaging an area including at least an index and an information image on a work reference object;
A vision system, comprising: an image processing unit that performs image processing on a captured image captured by the imaging unit to acquire index position information and index position shift information. A robot apparatus using the work reference object according to any one of claims 2 to 9,
A holding unit for holding a work reference object at a position where the robot arm can perform processing;
An imaging unit capable of imaging an area including at least an index and an information image on the work reference object held by the holding unit;
An image processing unit that performs image processing on a captured image captured by the imaging unit, and acquires position information of the index and position displacement information of the index;
A robot apparatus, comprising: an adjustment unit that adjusts a coordinate system used for controlling the robot arm based on position information of the index acquired by the image processing unit and position displacement information of the index. A reference object for work used in a robot apparatus having a robot arm,
An index that can be optically detected to adjust the target position of the work on the work reference object by the robot arm;
A work reference object, comprising: an information image that can be optically read to obtain positional displacement information of an index with respect to the work reference object. A robot apparatus using the work reference object according to claim 15,
A holding unit for holding the work reference object at a position where the robot arm can perform processing;
An imaging unit capable of imaging an area including at least an index and an information image on the work reference object held by the holding unit;
An image processing unit that performs image processing on a captured image captured by the imaging unit, and acquires position information of the index and position displacement information of the index;
An adjustment unit that adjusts a target position of the work on the reference object for work by the robot arm based on the position information of the index acquired by the image processing unit and the position shift information of the index. Robot device. An index member used in a robot apparatus having a robot arm and a working reference object,
An index unit capable of optically detecting and adjusting a coordinate system used for controlling the robot arm;
An information image recording unit capable of optically reading and recording an information image capable of acquiring the positional deviation information of the index with respect to the work reference object independently of the index unit;
An index member comprising: a fitting portion that fits into a hole portion of the working reference object.

Images