• 31 AUG: Nathan made an excellent presentation of the Stanley Overview paper, and we had a lively discussion. We identified several themes of the paper:
  • Testing is of tantamount importance
  • Simplicity/robustness
  • Software (not hardware) is the real problem
  • Maintain a "fallback" capability
  • Use learning in a pinpoint fashion (to learn a threshold for an algorithm, not a whole mapping)
  • Start from sound theory, add then add the second-order practical bits to make it work
  Kati proposed the Three Laws of Autonomous Vehicles. (cf. Asimov's Laws of Robotics).
• 29 AUG: We went over the course info. on this website, the "AI: State of the Art" handout below, and watched several video clips from the Nova sites linked below:
  • The Great Robot Race: A Hard Problem, Stanley
  • Meet the Teams: Ghostrider, Team DAD, ENSCO, Red Team, Virginia Tech
  • Outtakes: Balancing Act, Desert Sandstorm
  To do for Thursday:
  • Watch the rest of the "Meet the Teams" videos (30s each)
  • Watch the Outtakes video "DAD's Big Day"
  • Read through the presentation "What Robots See"
  • Read the linked Stanley Overview paper below

---

Created: 2006-12-02. Last Modified: 2006-12-02.