Welcome to Robotics! Spring 2007 Sarah Lawrence College Professor Jim Marshall
Robots Have Many Uses Welding Assembly Pumping gas Packaging Eating cars Dancing
What Is Robotics About? ● Robot: A physical system that autonomously senses the environment and acts in it ● Engineering and industrial applications ● AI and cognitive science research ● Both hardware and software
Some Early Robots Shakey (SRI, 1969) The Stanford Cart (1977)
Cartland
Behavior-Based Robotics Genghis (Rodney Brooks, MIT) Hannibal
The Cog Project (Rodney Brooks, MIT)
The Cog Project
Kismet (Rodney Brooks, Cynthia Breazeal, MIT)
Roomba Vacuuming Robot (Rodney Brooks, iRobot Corp.)
Handyboard/LEGO Robots
Other Popular Platforms Khepera II Koala Pioneer 3
Hemisson AIBO ASIMO Coco Ants Wheelsley
AIBOs Playing Soccer
Outline of the Course ● Web page (up by Monday) obotics ● Software tools we’ll use this semester – Linux – Emacs – Python – Pyro – Interactive C
Outline of the Course ● Introduction – Overview of AI and robotics research – Getting comfortable with Linux, Emacs, and Python – Robot morphology, sensors, and actuators ● Part I: Direct Control – Midterm project demos ● Part II: Learning and Development – Final project presentations
Part I: Direct Control ● Reactive control – Braitenberg vehicles – Handyboards, LEGOs, and Interactive C ● State-based control – Finite-state machines – Subsumption architecture ● Vision ● Midterm project demos
Part II: Learning and Adaptation ● Neural network learning – Pattern association and backpropagation – Simple recurrent networks – Self-organizing maps ● Evolutionary robotics – Genetic algorithms – Evolving neural networks with GAs ● Developmental robotics research – Readings and discussion ● Final project presentations
Project I: Possible Ideas ● Enable a robot to recognize where it is while wandering around trying to get to class (or back to its home base in the lab) ● Clear objects out of the environment ● Collect objects in the environment and return them to a target location ● Do something with vision and/or gripper using the Khepera ● Build new sensors for the Handyboard
Project I: Possible Ideas ● Explore sensor-limited and computation- limited robotics by using Lego RCX bricks ● AIBO – follow the ball – kick a ball into a goal – cooperate in accomplishing some task – play mastermind or some other visual game ● Treasure hunt: visually follow a trail of colored paper or other cues
Project I: Possible Ideas ● Investigate Player/Stage/Gazebo robot simulators ● Investigate Monte Carlo localization and other navigation algorithms ● AAAI or Beyond Botball competitions ● Investigate other robotic platforms (Hemisson, XBC, e-Puck, ER1, etc., depending on hardware availability) ● Other possibilities -- be creative!
Assignment 1 ● Log into your Linux account (once we get them set up!) ● Try out Emacs and go through the tutorial ● Start reading Python tutorials on the web and play around with Python ● Start reading Vehicles ● Check course web page later for details ● See you Monday!