From February 25th to March 9th, I’ve completed a number of tasks so that I could be prepared to build the robot after the spring break.
First, I’ve created a CAD model of our robot.
This will be the guidance to building our robot base and was added as figures in the design review report.
Then, my teammate Hirani and I went to Roboclub to source the materials for robot base. We found out that there were lots of aluminium extrusions and acrylic boards provided for us. However, we could not locate the connectors between aluminium extrusions and acrylic boards, so we decided to purchase them later.
In addition, we finalized the dimensions and mechanisms of conveyor belts together during class. As can be seen in the above pictures, we will have two conveyor belts: a flat, shorter belt that carries the trash items to the other longer, slanted belt which dumps them. We have not confirmed the exact materials we will use for the conveyor belts, but the most probable one would be pvc boards, which are highly durable. Please refer to the links for further explorations:
https://www.youtube.com/watch?v=UsF5Isjdgw4
https://www.youtube.com/watch?v=8Vnoso8rhaw
https://www.youtube.com/watch?v=VSJW71yo2oY
Also, I’ve worked on path finding/planning algorithm and come up with an algorithm. The reason for having this is to want to make sure that the robot efficiently moves around, avoiding the locations it has been to. The algorithm is explained in this google document.
https://docs.google.com/document/d/1P4U51VM7-ELiR2dowG1Jl_a368t6pJS8dFdUspEPo6w/edit?usp=sharing
However, we found some issues during discussions with our current camera. As it is a stereo camera, it cannot support many options that use map localization and depth search. Therefore, I decided to change the camera to Intel Realsense Depth camera (https://www.youtube.com/watch?v=mNTi1osPSIA -this is the video that inspired me to use Intel Realsense camera) and did some more research on it with regard to its compatibility with Jetson Orin and functionalities. For now, using the library AprilTag looks promising as can be seen in the gifs according to this link:
https://nvidia-isaac-ros.github.io/repositories_and_packages/isaac_ros_apriltag/index.html
Finally, I have worked on the design review report until Friday before spring break.
I am slightly behind the schedule, as I was supposed to set up the camera and finish the path planning algorithm before spring break. However, once I receive the Intel Realsense camera that I ordered, it will not be a hassle to complete the above tasks in a timely manner, since the reason for delaying was the incompatibility of our current camera with the tools mostly used for path planning algorithm.
Therefore, my action items will be the following:
- Build a CAD model for our roller with actual dimensions
- Finalize path planning algorithm with the camera setup
- Start building the robot base with the sourced materials
0 Comments