CS287 Advanced Robotics

Course Description

This is a graduate course on robotics for computers scientists and those with an engineering or physics background. Traditionally, the goal of robotics (anthropomorphism) has been to build machines with human-like dexterity and/or intelligence and which function with minimal human intervention. But recent developments suggest that robotics should encompass a much larger set of goals, and that the most dramatic growth will come in areas that depart radically from the anthropomorphic ideal. This offering of the course will attempt to plot these new directions, and extrapolate to the state of robotics 10-20 years hence.

Some background in computational geometry and control theory is essential and rather than assume this, the first few weeks will cover the material we need. Topics include data structures, sweepline algorithms, voronoi diagrams, convex hulls and linear programming, duality, epsilon-nets, dynamics and model-based control. The core topics that follow are (i) telerobotics and telesurgery, (ii) agents and model-based vs. behavior-based architectures (iii) RISC and minimalist designs for sensors and manipulators (iv) micro-robotics, silicon motors, grippers, motion arrays and sensors (v) nano-technology, SPMs, molecular pistons, Fullerenes, "lego" proteins.

Administrative Details

Lecture: MW 10:30-12 in 505 Soda
Prerequisites: Some knowledge of computer algorithms and mathematical background equivalent to M53-54.
Credit: 3 units
Instructor: John Canny, jfc@cs, 529 Soda, 642-9955
Office hours: Tues 2-3, Weds 3-4
Course Secretary: ???, liona@cs, 719 Soda Hall, 642-9575
  • bi-Weekly Homeworks 45%
  • 2 midterms 25%
  • Final project30%
  • Course Text:
    The course material will be covered in a reader available next week and in online notes.

    Back to the CS287 home page.