Humanoid robots such as Data, Terminator and Rosie are more than just anthroporphic machines. We envision these robots having significant computational resources that enable them to control their minds and bodies.
In this class we will study what it means to develop and evaluate autonomous and semi-autonomous decision making algorithms for robots that reach towards human level mental and physical capabilities.
Presently, we have two human size humanoid robots in the Humanoid Robotics Lab at Georgia Tech. We will apply our algorithms to these platform to validate these and other intelligent actions performed by the robot such as walking and general manipulation.
Members of this class will operate as a team, operating in both simulation and on real robot hardware to maximize the autonomous capabilities of a humanoid robot and simultaneously learn about existing research and techniques that can be brought to bear to solve these challenging problems.
Students will form groups of two or three according to their skills and interests. Each group will focus on elements of algorithms, software and hardware development. Specific topics and responsibilities will change in accordance with progress made by the groups. Furthermore, collaboration between groups will be strongly encouraged.
Each student will be expected to contribute to the development of the robot platforms. The student will keep an online journal of results, findings, challenges and solutions. By the end of the semester, the instructor expects a detailed account of the work performed with pictures, videos, equations and anything else that shows progress on the project. This will benefit not only the students themselves, but all members of the class and the research community.
Each week will be split into two sections. On tuesdays, we will all meet to discuss results from the previous week and decide directions for the coming week. On thursdays, students will give in-depth presentations of recent work in humanoid robotics and lead class discussion on the consequences and relevance of existing research.
Grading will be based on the contribution to the development of our semi-autonomous humanoid robot, a presentation on external research, and a group presentation on progress made towards the goal of robot construction. Contributions will be evaluated based on effort and thoroughness of the online journal. The breakdown is as follows:
50% Online Journal
25% Individual Presentation & Discussion
25% Group Presentation on Results
The course has no official prerequisites. Some knowledge of programming will be assumed, preferably in C, C++ or Matlab as well as some background in Linear Algebra. Additional knowledge in Robotics, Control, Perception and AI will be very helpful.
We welcome students from all engineering majors such as EE, ME, ECE and CS. This course is highly interdisciplinary and will give students the opportunity to explore applications of their previous studies and acquire familiarity with other fields.