Over the course of the past 10 weeks, I have been working on a robot named TrussBot. In simplest terms, it is a robot constructed of only tetrahedra whose primary purpose will be to climb poles. The larger implications of our design are to make robot design modular, personalizable, intuitive, and accessible.
Our team consists of three members, each responsible for one of the three following arenas: actuation, fabrication, and simulation. I have been responsible for the simulation of the TrussBot. Using MATLAB, the code I have written aims to accurately display and predict the motion of the full TrussBot with specific motor parameters and initial configurations.
I was able to build a general shell for the classes and functions needed to accurately simulate the movement of the TrussBot. There were a few bumps and abandoned paths along the way. For example, one of my initial attempts at a collision check involved algorithms that I did not have any prior exposure with, and involved some math that I was not familiar with either. These algorithms ended up being a bit too difficult for me to learn and debug on my own, so I eventually went with a slightly more primitive approach. Currently, the new collision checks I wrote are not noticeably slower for the small tests I have been running. There may be room for more optimization in the future.
Though I had previously worked in the Sung Robotics Lab in the spring semester, I was working on a project independently (as part of another program) and my work then called for a much smaller time commitment. Coming into the lab and fully immersing myself in a work environment with other students pursuing research was very rewarding. Though we were all working in a robotics lab, the students came from all different disciplines, schools, and years of experience. Everyone had something to learn from everyone else.
I left this summer with a much stronger skillset for problem solving, team development, and computational geometry.