We finished the pick and place task really easily with the UF gripper. This allowed us the space and time to explore making our own end-effector.
The 'Peg in Hole' task was the most challenging for our team. We faced the challenge of having an offset when placing the peg, as seen in the fail clip of the video. Even after taking a second photo for recalibration, this offset was consistently a few cm off in the +x direction. Therefore, we hardcoded the offset correction, which worked.
Due to time limitations, we chose to show how the dual-end effector could potentially work with the software algorithm instead of actually integrating it. In the video demonstrations with the custom dual-end effector, we are using manual control to show that the dual-end effector is capable of picking and placing all the objects. However, to integrate the custom dual-end effector, we would have to override the code architecture that looks for the UF gripper. Furthermore, in order for ROS2 to communicate with the Arduino, which controls the actuators of the end-effector, we would have to use micro-ROS and ROS-serial. Finally, the computer is on the opposite end of the lab from the robot arm. Therefore, we would have to custom-make really long wires.
Overall, we faced time limitations that prevented us from doing the full implementation of the dual-end effector as we intended. However, we are fully confident that the combination of the software algorithm and functioning hardware would achieve the objectives. Relevant online tutorials are included in the Resources section for reference.