The FANUC robot is a LRMate 6 axis robotic arm with R30iA Mate controller. The robot can be controlled using a teaching pendant. THe teaching pendant is used to increment each joint angle. The speed of the robot can also be changed here.
Control Block diagram MATLAB
Visual Basic
PCDK libraries
Robot Controller
MATLAB is used to calculate or generate the coordinates of the end effector
The coordinates are then transferred to VB where the program to move the robot is written.
The compilation of the program uses the PCDK libraries to call the required functions.
The program is then transferred to the robot controller, which moves the robot.
Form And Code
Code's Structure First we made buttons on the form and the by defining some functions for those buttons we tried to program the robots motion. Outlines: 1) Through IP gets Connected 2) Joint angles are determined 3) Gripper position is determined 4) End effector is moved
We have defined a new user frame. According to this user frame we want to move the robot. Also we want to take images and calibrate them. The coordinates of this user frame relative to world frame and joint position are in this picture:
Define your desired User Frame on Teach Pendant
iRvision
iRvision: By typing the "IP"(192.168.1.18) of the robot controller on the address bar we can get access to the iRvision software:
Put the IP in the address bar
First of all we have to calibrate the images that we get from the camera by starting a new calibration procedure: This button is for creation of Camera Calibration tools, because of software issues, it has no specific shape on it.
Created as a new Calibration Tool
With the sheet that belongs to the calibration method we calibrate the images (note that robot's gripper has to be sat in the origin of the user frame when the procedure is going on). Also note to the data and combo bar's adjustments:
Read the grade spacing from here
Focal distance is an important parameter, which can be found from Fujinon's catalogue. It has to be right to give us correct "Z" height.
Read it from Fujinon’s Catalogue
Note to the camera and application frame(user frame)'s positions: These data are generated after setting the setup (you can always check for their validity)
The height of Camera relative to the Calibration Grid (Surface)
Calibration gives the coordinate of points on the screen:
Coordinates of Calibration Points
Points coordinations bar from a closer sight:
Y Origin
X
Don't forget to save the calibration procedure after completion of calibrating procedure. After removing the calibration sheet, we still can see the calibrated points are on the screen:
Save the Calibration after you are done with it.
From "Vision Process tools" section open up a new "2D single view locator". We named it "FRIDAY" as you can see the name in the picture. In the following slides we are going to adjust the adjustments to get the desired locating tool to detect the position of objects below the camera relative to Cal Grid: This button is for creation of Vision Process tools. Because of software issues, it has no specific shape on it.
Created as 2D single view locator process tool
In the beginning there is one GPM locator that can only be adjusted to detect one specific object, you can see the adjustment toolbar in the next slide for this specific object among the other objects.
Red box search window
After adjusting the search window press OK
Set the search window
One should use the application for each button from the manual. In brief, three buttons have the most application:
1)Teach Pattern 2)Set Origin 3)Set Search Window
After setting the adjustments we can detect the specific objects among the other objects; by adjusting the contrast we can detect the other 3 similar objects.
Not detected because it’s out of the search window
Finding Model ID 1
Not detected because of low contrast
Here the contrast has been adjusted and the seven similar objects have been detected by the process. Note that the 8th object is out of the blue search window which was adjusted in the previous steps, so, it has not been detected.
Finding Model ID 1
Here we have three kinds of objects, in the following slides we are going to detect each one of them, here objects model id 1 have been detected:
You can add new GPM (geometric pattern match) locator tools for each new object. Then, just adjust the pattern, origin and the search window if it is necessary.
The new object is detected according to its GPM locator:
Finding Model ID 2
GPM locator adjustment and the 3rd detected object in the image: Finding Model ID 3
After everything is done with the GPM locators we can detect all the objects together:
Each item has its own number on the screen; here we have eight objects of model id 1, one of model id 2 and one of model id 3:
Use the proper calibration (check slides 12-18)
The coordinate of items relative to the calibration grid frame and model id numbers are sorted in the table below iRvision's screen:
Example: In this random image, we want to detect objects of model id 1 among the other objects:
Using the FTP connection between PC and robot's controller, we can generate this picture:
Not detected because it’s out of the search window
Detected model id 1 objects among the other ones; compare the pictures:
We can open the generated picture with MATLAB and using the image processing commands detect the origin of each cross:
The directory in which the FTP generated picture is.
Pixel Coordinates of 7th detected object’s middle cross
Origin of the cross has been detected, and a dot has been sat up there to show that we can have the origin coordinates of each detected items. In this picture you see the dot points in the middle of crosses:
Dots in the middle of the crosses to show that we know their exact pixel coordinates
The future goals are: 1) Making difference among the crosses for different objects. 2) Transferring the pixel coordinates to Matlab 2-D. 3) Moving the gripper of the robot to the top of the object and picking it up and then, putting it in a desired location.
