CS225B Robot Programming Laboratory

1
CS225B Robot Programming Laboratory

• Email konolige_at_ai.sri.com, gerkey_at_ai.sri.com
• TA Phil Fong fongpwf_at_stanford.edu
• Website cs225b.stanford.edu
• Mailing list cs225B
• Projects done on robots in Lab
• Lab itself in B30/32
• Keys from Gates 176
• Office hours before class, evenings before
  projects due

2
Class Composition

• Target size is 25 students
• Priority
• CS grad
• CS undergrad
• Others
• Groups of 2-3 people for projects
• Try to get at least one CS person / good
  programmer per group
• Groups formed by next Tuesday
• Choose a group name, email me and TA

3
Course format

• Lectures, alternating between theory and practice
• Requirements
• Linear algebra and trigonometry (2D transforms)
• Probability theory helpful, but we will learn all
  we need
• C, Python, Java, etc. -- programming ability
• Linux OS helpful
• Texts
• Thrun, Burgard, Fox. Probabilistic
  Robotics.http//www.probabilistic-robotics.org
• LaValle.  Planning Algorithms.  Available online
  http//msl.cs.uiuc.edu/planning
• Papers and handouts
• Readings expected to be done by the day they are
  listed in the syllabus (20 of the grade)

4
Projects

• 3 Projects 1 Final Project
• 50 30 of grade
• Projects due every 2.5 weeks
• Demonstration class projects running on the
  robots
• Laboratory in B30/32
• 3 Linux desktop computers
• 3 Erratic robots, wireless access
• Simulator

5
Course Goals

• Learn about the autonomous robotics field
• Investigate the main methods needed to make
  robots move intelligently
• Hands-on projects to implement the methods

6
Course Outline

• Mobile Robot Motion and Architectures
• How do mobile robots move? How can we program
  them to perform tasks?
• Player/Stage architecture
• Localization
• Markov localization
• Probabilistic representation of uncertain
  movement
• Planning
• Metric and topological maps
• Wavefront methods
• Random tree methods
• Final Project
• Dynamic sensing (vision or LRF)
• Competitive environment

7
Final Project

• Task Program the robot to play one-on-one robot
  soccer
• System AmigoBot with laptop, color camera,
  sonars
• Abilities
• 1. Stay localized
• 2. Stay out of opponents goal
• 3. Locate the ball with laser
• 4. Movement to kick the ball
• 5. Movement to block the ball
• 6. Strategy

..\videos\auto-skate.avi