Devank Status Report for April 22, 2023

2The status report covers the last 2 weeks:

Week 1: April 8th to April 15th
This was the week of the carnival at CMU so I could not get a lot done because of the same. I made some changes to the code base, where it would aim for the outer edge of the pockets instead of the centers making sure the shots actually get made without colliding with certain walls. Moreover, we ran into a problem where the CV would pick up the projection of the balls as real balls and to overcome this we decided to integrate a servo that would move a piece of wood over the projector. So I ordered the parts for the same as well as thus came up with a new design for the frame.

Week 2: April 15th to April 22nd

This week was the majority of my work. The work was mostly divided into 2 parts.

The first of these was the Servo idea mentioned above. This week Jimmy and I worked on integrating that into the frame. Currently, two buttons (one for solids, one for stripes,) are connected to the breadboard at the back of the longest leg of the frame. When a user presses one of these, The servo is signaled by an Arduino to move and cover up the lens. Once the lens is covered, A signal is sent to Jimmy’s system which achieves 2 objectives: Tell the software what kind of ball to aim for, and to take a picture. Once the picture is taken and the shot calculation is done, The python code signals the Arduino back and then the servo is moved back into its resting place, displaying the optimal shot and ready for the next button press. I then worked on making the entire wiring and breadboard look pretty and not a mangled mess which it originally was, which took a lot more time than expected. The two buttons are now placed on opposing ends of the pool table and all of the electronics are hidden away from the user, with the servo being placed right under the projector for obvious reasons.

The second of these was the alignment of the projector with the frame. Because of the offset of the projector, it has been a real pain to ensure proper alignment of everything when being projected. This involved more woodcutting and software image rotation to ensure that the projector displays a non-tilted image on the table. We used the PIL API and OpenCV to get this done from the software perspective and just woodworking from the hardware perspective.

In terms of my next tasks, we have our final presentation on Monday/Wednesday which I am doing and thus we have to make the slides for that and I have to prepare for that. We have to get our report and poster ready in time for our demos and I still have a bit more of projector alignment to fix. Tomorrow I am planning on integrating the smart cue stick into the system with Justin so that we can have all things finally integrated and move onto removing the small bugs and caveat’s over the next two weeks. The end to end latency of the system right now is somewhere around 2 seconds (button press to projection) and will become a lot more consistent when we have constant lighting which shall be in place by the upcoming Monday.