Propose your final project, then implement it and write a short report documenting your work. Your group will also present and demonstrate your final project during the final exam period.
Write up a short project proposal as a wiki page. The proposal should be a clear description of what you want the robot to be able todo. Ideally, the task should incorporate vision, motion, and sensing. The task does not have to involve human interaction. It could involve identifying and manipulating objects in the environment, or simply navigating the environment using some kind of localization, with goals set by some kind of interface (e.g. a web page or remote operator). Give the page the label cs363s17proposal.
My suggestion is to pick one facet of the project that you want to showcase. Some realistic possibilities include:
- Robust human-robot interaction for a particular task (like greeting people coming into the building).
- Some type of outdoor navigation, possibly following sidewalks.
- Some type of RL type learning, best with discrete states and actions.
- Vision-based landmark localization, possibly using posters/paintings.
- Remote operation of the robot, preferably coupled with some kind of localization.
- Machine learning for perception or control.
- Robot hide and seek (visual recognition and hiding/searching strategies.)
- Modeling a dog's behavior.
Your proposal should have a diagram or clear description of each separate module and how they will communicate (e.g. vision module, nav module, control module ).
Your proposal should include weekly goals to demonstrate during lab times on April 27th and May 4th.
In addition to the report for the project, you will have 20min to present your project to the class during the final exam period.
The writeup for your project should be in the form of a 6-8 page conference paper. You can use the author guidelines for an IEEE conference paper (here).
The paper should include the following sections.
- Short introduction and problem description
- Related work
- A description of the physical hardware
- A description of the software, including algorithms and architecture
- Results and experiments
- Summary and conclusions
Send your final paper to the professor as a pdf. You may also create an associated zip file or web page, referenced in the paper, in which you can put additional images and videos.
Finally, put a full copy of your final project code directory on the Courses handin folder. The easiest way is to zip the directory, copy it to a mac, and then copy it over to Courses.