JPH04310360A - Burr removing robot control method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to an industrial robot that performs work such as deburring machined products or castings in accordance with command signals, and in particular, the present invention relates to an industrial robot that performs work such as deburring machined products or castings in accordance with command signals. Concerning a robot control method that removes burrs with an accuracy equal to or higher than that of

0002

[Prior Art] Conventionally, deburring is a typical work carried out in a harsh environment, so automation using industrial robots has been progressing. December 1988
Discussed in May, pp. 52-56.

0003

[Problems to be Solved by the Invention] In the above-mentioned conventional technology, a robot is taught in advance a predetermined path where burrs may occur on the target workpiece, and a force sensor is used to detect burrs on the path. Based on the information, compliance constants (eg, spring constants, viscosity constants, etc.) are changed on software to perform force feedback control. However, although it is fully applicable to removing burrs from relatively small and simple shapes such as machined products, the conventional technology has a finishing accuracy of ±0 when removing large burrs (height 5 mm, width 2 mm) on the surface of castings etc. It is difficult to keep the thickness to less than .3mm, and the length of weirs, etc. that occur on the milled surface, which is unique to castings, is 10 to 20 mm.
It takes a long time (several tens of minutes) to teach the route because it is irregular on the order of mm and differs from work to work.

An object of the present invention is to provide feedback control means based on burr shape recognition for efficiently removing relatively large casting burrs.

[0005]

[Means for Solving the Problems] In order to achieve the above object, the present invention equips a robot with a visual sensor that recognizes the burr shape, and uses this burr shape information, tool reaction force from a force sensor, and rotation of the tool. The robot operation information is generated from the number and feed rate, and the robot changes the deburring conditions and processing conditions in real time according to the shape of the burr while observing the burr, and adjusts the tool pressure, feed rate, and rotation. Control the number, etc. In addition, the deburring path is created in advance by CA of the workpiece.
D (Computer Aided Design
) Extract the ridgeline where the burr occurs from the data, and then use the NC (
NuwmericControl) data,
This can be achieved by teaching.

[0006]

[Operation] First, the controller of the deburring robot has
A workstation connected to an external local area network sends burr location identification data converted by software that extracts ridgelines (line segments) from workpiece design drawing information. At the same time, data on the burr occurrence position on the milled surface, which was created separately, is sent.

On the other hand, a visual sensor attached to the robot's wrist sends pattern data of the burr shape (burr width and height) to the controller while monitoring along the ridgeline of the workpiece. Based on this burr shape, the controller
Make a deburring plan. For example, if the burr shape is large, the robot drive information can be generated by selecting the number of trials, type of tool, and processing conditions to repeat the operation several times.
Commands deburring operation.

[0008]

[Embodiment] An embodiment of the present invention will be described below with reference to FIG. The deburring robot controller 1 controls the rotational speed or speed of the tool 6 held by the robot 3 and the feed rate of the mechanism that presses the entire tool 6 against the workpiece 7. Further, the controller is sent NC data obtained by extracting the ridgeline (line segment) of the burr occurrence location from the design drawing information created by the workstation 13 connected to the external local area network 12. Along with that,
Information on the location of burrs on the milled surface of the casting is sent. The controller 1 determines the moving distance (route) of the robot 3 from the coordinate origin based on the NC data and position data.
is calculated and an operation command for the tool 6 is output.

On the other hand, a visual sensor 5 attached to the wrist of the robot 3 detects the size of a burr 8 on the ridgeline of the workpiece 7. To detect burrs, as shown in Figure 2, a laser beam is scanned in a direction perpendicular to the ridgeline, and at each scan, for example, the cross-sectional shape of the burr height h and width d (or area S) for burr positions A and B is detected. are detected as in patterns 1 and 2. The image processing unit 10 determines the size of the burr (h, d, S, etc.) using the casting surface of each pattern as a reference, and sends burr shape information to the controller 1 and the fuzzy controller 11. Next, the fuzzy controller 11 performs fuzzy inference from the tool reaction force of the force sensor and the burr shape information,
Target values for the tool rotation speed R and the feed speed V are calculated and output. The fuzzy inference rules are determined by the tool reaction force and the degree of burr shape (large, medium, small), for example, as shown in FIG. An example of a rule is, ``If the burr pattern is large and the tool reaction force is large, the tool rotation speed should be much faster and the tool feed rate should be moderate.'' The target value of rotation speed and feed rate can be shown quantitatively. The method is to synthesize membership functions that indicate the likelihood determined by the goodness of fit from each membership function, and easily obtain the centroid of the membership functions.

FIG. 4 shows membership functions that characterize the burr shape. Further, FIG. 5 shows the membership function of tool rotation speed R and feed rate V depending on the burr shape. The fuzzy controller 11 can be easily realized by software in the controller 1 or commercially available hardware integrated into a single chip, and since the inference rules can be easily rewritten in memory, it can be easily implemented by a skilled worker. Deburring know-how can be easily incorporated into robots in the form of rules. The above operations can be achieved in tens to hundreds of microseconds per rule by using a hardware fuzzy controller, so it can be achieved by sequentially and repeatedly controlling the tool for visually recognized burr coordinates. You can remove burrs from the workpiece in a short amount of time. Further, the mounting position from the visual sensor does not necessarily have to be on the same arm of the same robot.

[0011]

[Effects of the Invention] According to the present invention, it is possible to perform relatively large casting deburring work visually, systematically, and efficiently in real time, eliminating the need for heavy-duty manual work and eliminating the need for manual work. This reduces the amount of time required for deburring work, and allows deburring to be performed with accuracy equal to or higher than that of a skilled worker.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention;

[Figure 2]
An explanatory diagram of an example of a detected burr pattern,

[Figure 3] Explanatory diagram of fuzzy inference rules,

[Fig. 4] Explanatory diagram of membership function of burr shape,

FIG. 5 is an explanatory diagram of a membership function of tool rotation speed R and feed rate V.

[Explanation of symbols]

1... Controller, 2... Motor driver, 3... Robot, 4... Force sensor, 5... Visual sensor, 6... Tool, 7... Workpiece, 8... Burr, 9... Mechanism monitor interface, 10
...Image processing unit, 11...Fuzzy controller, 1
2...Local area network, 13...Workstation, 14...Link adapter.

Claims (1)

[Claims] [Claim 1] A visual sensor and an image processing device that detect the burr shape are provided, the reaction force of the tool is detected by a force sensor, and the rotational speed and feed rate of the tool are determined from information on the burr width and its height or area. A method of controlling a deburring robot characterized by controlling the robot using fuzzy reasoning.