Computer Science 154
Robotics, Spring 2010
Assignments/Projects



There are two main components to the coursework in CS 154: larger, robot-based "lab projects" and several short assignments with written and/or programming questions.

Assignments




Lab Project Possibilities

I would encourage everyone to form teams of about 3 people - you're welcome to propose a semester long project, or to try something out for the first half of the term and then switch (teams and/or projects) after spring break.

Here is a summary of the options presented in the first lecture's slides. They're organized first by platform, and then by possible tasks to pursue with that platform. Do keep in mind that these can be for mixing/matching, for inspiration for a completely different project, or can be taken as they stand.
  • Default Tasks

    The default platform is the iRobot Create, although any of the prebuilt platforms will work. The Create is a very adaptable robot for indoor tasks. The default set of tasks will guide you in creating a "silicon Mudder", that is, a robot that can navigate intelligently among the halls of the Libra complex (or another similar environment, e.g., the dorms). Those tasks will guide you through building some basic capabilities during the first half of the term:

    • Getting started with the software
    • Wandering, sensing, and state-machine control
    • Localization: finding oneself on a map, even with noisy sensors
    • Mapping: creating a map, even with noisy sensors

    These are also meant as foundation for additional projects you might pursue in the second half of the term.


  • Other Create/Roomba projects

    • using the vacuum - Roomba Pac-Man
    • building your own iRobot ConnectR - telepresent robot
    • multiple-robot cooperation or competition -- we have lots of roombas!)
    • gaming with (probably not against) robots...
    • iRobot's hamsterbot (cool!)
    • delivery-bot, perhaps anywhere within the libra complex passages
    • unusual mapping algorithms: for example, using the WiFi signal strength
    • The AAAI robotics exhibition: in Atlanta this summer


  • The Lego NXT (we have a couple of these)

    • spatial reasoning labs, a.k.a. the silicon Mudder
    • I'd love to be part of the first mapping algorithms implemented on a Lego robot... :-)
      • competition in Pasadena this summer
    • tag, hide-and-seek, competitetive games
    • fire extinguishing in a maze (or, at least, candle-extinguishing)
    • if it can be done at all, it can be done with Lego! Prove this statement true.


  • The Chiaras

    • this is not a cute platform: its' fairly formidable-looking. You can check out a video from last summer at this YouTube link
    • This robot would not only support the spatial reasoning tasks (the default), but would be an excellent platform to play a game, for example, against a human. There is a chess-playing competition in Atlanta this summer but other games/demos with these two robots would be wonderful! (Nim, anyone?)
    • It offers as much more more manipulation capability as the other platforms -- as a result, it could also undertake clean-up or even construction tasks.


  • The Scribbler/Fluke combination

    • this is a cute platform, more for tabletops than human-scale environments, but it would certainly support the default silicon Mudder spatial-reasoning tasks
    • these could also be used for investigating multi-robot coordination algorithms


  • Sony's AIBO dog

    • spatial reasoning labs, a.k.a. the silicon Mudder (again, the default)
    • a robot soccer player
    • teach our dog new tricks -- machine learning / neural nets applications


  • The Evolution ER1

    • spatial reasoning labs, a.k.a. the silicon Mudder (again, the default)
    • AAAI scavenger hunt competition
    • my goal has long been to have a robot that can call and ride the elevators
    • this is a modular kit and can probably handle almost any indoor task...
    • This was 2009's "Ultimate party bot"


  • PowerWheels Robots - Mudd should have the first BarbieJeep tour guide!

    • spatial reasoning labs, a.k.a. the silicon Mudder -- but in this case, the spatial reasoning would be done outside, rather than in the halls
    • the Mini Grand Challenge, a.k.a. the Urban Challenge for Barbie Jeeps
    • the Intelligent Ground Vehicle Competition
    • a campus-wide tour guide
    • other outdoor-robot fun :-) delivery? pick-up?


  • Competitions / Simulators / Other ideas...

    • a robot that plays a game on the Wii - updating the DuckHunt robot for '09
      This would probably involve creating a small robot arm and mounting a Wiimote to it.
    • even more recursive: a robot that controls another robot with a Wiimote !
    • a robotic tie-in to an existing project you might be working on
    • design/build/program your own, but I have found that one term is really not enough time to build off-the-ground platforms (such as aerial or underwater robots) -- and then undertake algorithmic or computational experiements with them. But if those are your passion, I'm happy to talk!


Course Grading

Each assignment is graded based on its particular tasks and normalized out of 100 points. However, the overall assignment of grades follows the customary, general headings as follows:

  • A (90-100) for outstanding work
  • B (80-90) for solid work
  • C (65-80) for dialed-in work
  • D (50-65) what happened?
  • F you didn't show up
with shades of distinction indicated by the usual pluses and minuses.

The weight that the assignments and projects contribute to a final grade breaks down as follows:

  • Assignment 1 and participation: 10% of the total
  • ~5 assignments, about every two weeks: 50% of the total
  • Lab/Final Project(s): 40% of the total

Collaboration and the Honor Code