CMRoboBits: Creating an Intelligent AIBO Robot Instructors: Prof. Manuela Veloso & Scott Lenser TAs: Douglas Vail & Nick Aiwazian Further Support: Dr. Paul Rybski, Sonia Chernova , Fall Computer Science Department Carnegie Mellon
Why This New Course? Our robot soccer experience – Long: since 1996 – Diversified: simulation, small-size, Sony 4-legged – Hard work, frustration, fun, struggle, success – A LOT learned on: Creating teams of complete intelligent robots.
Legged Robots Remarkable hardware - SONY AIBO robots
Intelligent Complete Robot Action Actuators Perception External World Sensors Cognition
Autonomy Perception – sensing, modeling of the world Cognition – behaviors, action selection, planning, learning – multi-robot coordination, teamwork – response to opponent, multi-agent learning Action – motion, navigation, obstacle avoidance
Perception: Vision Real-time and robust Effective calibration Colored blobs identified as objects Confidence computed
Cognition: Behaviors ScoreSearch Approach Recover not see ball next to ball not see ball timeout see ballnot next to ball
Action: Motion Four-legged walking Head motion Turning, kicking
What is This New Course? Beyond robot soccer CMRoboBits: – From CMPack, break into modular robot “bits” – Learn the CMRoboBits within a concrete robot AIBO – Commercially available to the public since June’02 – Robust hardware, modular software
The Syllabus Sensors Basic behaviors Motion Vision Localization Complex behaviors Multi-robot behaviors
The Evaluation Weekly homeworks – hands-on lab work with the AIBO robots – some also include written questions One final project Homeworks: 75% Final project: 25% Groups of 2 or 3 students: each student in the group must be fully responsible for each homework
Resources and Readings Readings and videos are available at: AIBOs are being used widely in RoboCup: The OpenR Web page has a lot of information: Tekkotsu is a serious effort to build an API for the AIBOs: