Final Project Design
The purpose of this week is to plan your final project, which will take up the remainder of the semester. The project should be something of interest to you and can be either the implementation of an important capability, such as localization, or the development of a robot system to engage in an activity, such as a robot greeter or photographer.
The end result of this week should be a specific, week-by-week plan of your project, set up as a wiki page. You may use any of the robots in the lab: the Magellans, the Darwins, or the ATRV. If you chose to use the Darwins, then you'll need to commit to probably two weeks of intense learning to understand how their control/walking software works.
The items you need to complete this week are the following.
- Build your group. You may have groups of up to four people. The list of people in your group must be given on the wiki page.
- Decide on a main problem statement for your project. Put this on your wiki page.
- Develop a week-by-week plan of action, similar to a Gantt chart. This should include an initial breakdown as to who will be responsible for what parts of the project. Your chart should be on the wiki page.
- Develop a list of resource materials, including at least 2-3 relevant technical papers. These should be listed on your project wiki page.
- Present your design and plan during class on Thursday, 11 April. These should be no more than 10 minutes.
If you spend some quality time designing and planning your project, it will make the following weeks much more successful. I'm happy to talk with your group about ideas, resources, and planning.
Some suggestions for the projects.
- Localization and navigation, given a map. Be able to set the robot down anywhere on Bob's 2nd, get it to localize itself and then move from one location to another. This should be set up as a standalone program that gets data from the laser/sonars, executes the localization, and sends jump messages to the navigation module to update the odometry.
- Simultaneous Localization and Mapping (SLAM). Set the robot down and let it both build a map and localize itself on the map. This is a challenging problem.
- A robot greeter / photographer / tourguide. Implement a robot system that is able to greet people when they come into the lobby and provide some type of service.
- A robot assistant (outside). The robot can greet people outside Bob's or Miller and offer to follow behind them, carrying something. Another option would be a tourguide that can talk about landmarks on the way.
- A soccer goalie. Program the robot to track a ball rolled in its direction. The robot tries to block it from getting past into a goal. This could be done with the Darwins or the Magellans.
- Robot hide and seek. Program one or more robots to play hide and seek, preferably with people. You can engineer both the people and the robots with recognizable targets. You may want to implement localization for this.
- Visual mapping/localization. Use a camera instead of the laser to implement just localization or the full SLAM. There are some interesting possibilities here with taking and storing large numbers of images of the space and searching those images efficiently.
No extensions this week.
Label your project wiki page with the label cs363s13project6