top of page

WEEK 3

SOLVING TECHNICAL ISSUES THROUGH 

COLLABORATION

business1.png

Firstly, let's consider the last week's plan for this week:

  • Work on the control of the UR3

  • Computer Vision algorithm to classify colour blocks

  • Work on visual servoing to locate a point in space.

For the task of the visual servoing, I  still had some problems with the simulation, so I did not finish it this week.  For the control of the UR3, I finally got the Matlab code to work with the real hardware, which I had the robot to move around from one particular point to another and had it to draw for me a circle. One of my classmate, Tim Yan Muk, helped me with the connection of the UR3, which was pretty not a  straight forward procedure for the first-time user like me. For the Computer Vision, it is the first requirement of the visual servoing for the UR3, so it was fortunate that I had it finished quite early than expected. The algorithm will capture the frames recorded from a webcam, applies some masking and thresholding techniques, to output to the screen the color of the blocks.

Here are some videos for demonstrating my works for the previous week:

 

 

After the first week, I got an invitation from Meet Patel to work with him in controlling the UR3 to perform some industrial fabrication with a 3D printed part. He was working on a multi-tool mounting part for the UR3, which included a drill, laser engraver, sanding tool and milling tool. He planned on combining his tool and my controller of the UR3 to make his progress more accurate and user customisable. We started working on the drilling part first because it was the easiest part to

To make the UR3 to move to the position that we wanted, we needed a new working with an origin which was based on the world origin of the robot. As I was the one who designed and controlled the UR, I had to design the base. I had the prototype done on Wednesday so that we could test it. It turned out the way. 

 

 

 

 

 

 

 

 

 

 

​

However, as we were using mechanical tools to work with a 3D printed part, we needed to make sure that that part would be positioned securely on the base, so we came up with a new design with a custom - made clamp. 

 

Final result.jpg
20190224_170557.jpg

Finally, with the new base ready, we were able to perform our first drilling test with the UR3. It turned out to be quite successful. The robot moved as planned and the hole was clean and smooth.

It was the third week of the studio, and I have not finished the visual servoing part for the UR3, so in next 4 days of the fourth week, before the final demonstration, I want to finish the following targets:

  • Try to finish the Visual Servoing simulation and derive the algorithm.

  • Get the Computer Vision to work and extract the pixel positions of the corners of the blocks.

  • Integrate those points with the Visual Servoing and implement the servoing part.

  • Combine everything I had to a big project

bottom of page