Syllabus

CS B659: Principles of Intelligent Robot Motion

Spring 2011

Instructor: Kris Hauser

Intelligent agents need to coordinate many degrees-of-freedom under complex operational constraints to achieve long-term tasks, to sense and react to disturbances in real-time, and to interact with human operators and other agents. Principles of Intelligent Robot Motion is a graduate seminar course covering models, theories, and algorithms for motion planning and control, with applications to robots, humans, intelligent vehicles, virtual characters, biological molecules, and smart medical devices.

Subjects will include:

Modeling and representation
3D transformations and geometry, forward and inverse kinematics, dynamics and simulation of multibody systems.
Motion planning
Configuration space, sample-based motion planning, nonholonomic systems, planning under uncertainty.
Introductory control theory
Feedback control and asymptotic stability, operational space and force control, optimal and model predictive control, Markov decision processes.

Sensing, estimating, and interpreting motion
Kalman filtering and particle filtering, Hidden Markov models, Inverse optimal control.

Course content will consist of lectures, readings, and interactive labs. Students will complete semester projects, alone or in small groups, on a topic of their choosing.

Prerequisites: linear algebra and programming experience. Multivariable calculus recommended.