This summer, I worked as part of MinDART (Minnesota Distributed Robotics Team) with Paul Rybski, Chris Flowers-Huebner, Esra Kadioglu, and Harini Veeraraghavan, all of whom are graduate students at the University of Minnesota. The minDART uses robots to implement a task of search and retrieval in which the goal is to locate, collect, and return targets to a home base. Our homogenous robots had previously performed these tasks independently, using localization without communication for target search and retrieval. This tactic, while effective, was not efficient. Through our work this summer, the robots now implicitly communicate with each other through the use of beacons and cameras. This communication decreases the time for localization as well as the recovery of targets.
The work that I did this summer was not what I had expected to do. I have worked with Lego Robots and the Handyboard before, and so I was fairly familiar with the environment for programming. In addition, I had a lot of experience with UNIX and LINUX environment. I had thought that I would be able to write code to help implement the changes that needed to be made. Instead, I ended up doing a lot of testing and implementing hardware designs. I knew coming into this work, from my past experiences with LEGO Robots, that a large part of any robotics project is maintaining and adapting the physical parts of the robots. However, I wish that this summer I would have had more experience working with the actual code for development.
My work this summer was basically work that anyone could have done. There was no real work that I did that required a computer science background. I am a very good solder-er now, based off my soldering the circuits and the CMUcams. I also learned some things about designing circuits, which was useful knowledge to know how to do, although not neccesarily for Computer Science. I think that one of the major problems was that the MinDART team needed to get the code working fairly quickly, and it would not have made sense time wise for me to get familiar with all of the code as well as doing some of the actual coding. In addition, part of the limited work that I did in terms of coding was my fault. I could have come up with my own research project this summer and worked on it with the robots. Dr. Gini gave me that opportunity. However, I chose to work with the MinDART team because I thought it would be beneficial to work in an environment that was already developed.
This summer, however, was not anywhere near a waste. I really think that I gained some valuable insight into graduate schools. Through my research and experiences at the University of Minnesota, I got to meet graduate students, who were very helpful in any questions that I had. I saw what a graduate school environment can be like. In addition, I got to meet Dr. Gini, who had given me some confidence about my abilities.
Furthermore, through my experiences at the research talk of Holly Yanco set up by Dr. Gini, and through the AAAI Conference, I have a better idea of what I would like to do with my computer science career. Before, I kind of fell into AI because that is what my experience has been in. I wasn't sure that I really liked it. However, now I am pretty sure that is what I would like to go to graduate school in. The field of AI is huge and has many different possibilities. In addition, there is a lot of room for cross-discipline studies, which is something that I am very interested in. I enjoyed my experience with the CRA-DMP primarily because it gave me ideas about what I want to do in the future, as well as providing me with numerous contacts in the academic world that I can call upon for future reference.