Course Description

This course is about robot algorithms that allow mobile robots to operate successfully in unknown and unstructured environments. Currently many of the most successful algorithms utilize probabilistic approaches that explicitly encode uncertainty, use these representations to assimilate new sensor data, and base their decisions both on the content and quality of information. The material covered in this class includes the most prevalent and successful models and techniques, specifically the various versions of synchronous localization and mapping (SLAM) and randomized algorithms for motion planning. Depending on class size you will be able to implement these algorithms on a physical robot and have it roam the hallways of the department!

Course Logistics

Format Course Prerequisites

Linear algebra, introductory probability and statistics, programming experience in C++ and a computational scripting language such as Matlab, Octave, or NumPy.

NOTE: This is a new class. If you would like to take it but have trouble registering because you haven't taken CSE 568 and/or are not in the CSE department please e-mail me. I can help you figure out if you have the prerequisite skills and get you registered.

Topics

Resources

This class will loosely follow the book: Probabilistic Robotics by Thrun, Burgard, and Fox.

Algorithms: SLAM, EKF-SLAM, GraphSLAM, FastSLAM, RRT, RRT*, PRM