COMP 790-099: Motion Planning in Real and Virtual Worlds

Spring 2010

Time: Tuesdays, Thursdays 3:30pm - 4:45pm

Location: SN 115

Instructor: Prof. Ron Alterovitz

Office hours: Tuesdays 4:45pm - 6:00pm or by appointment, 223 Sitterson Hall


Animating movies, automating laboratories, controlling medical devices, and many other tasks depend on planning the motions of real or virtual objects. Motion planning algorithms compute sequences of motions that achieve a particular goal, such as maneuvering a robot arm in an automation task, controlling an autonomous vehicle, or simulating a character's movement in a game or virtual environment. Developing motion planning algorithms is both technically broad and theoretically deep, raising a unique combination of questions in algorithm design, computational geometry, control theory, and robotics.

The course will begin by introducing the fundamentals of motion planning algorithms and then lead to discussions on current research and applications. The topics to be covered, which may be modified based on the interests of enrolled students, include:

Who should attend: Students with interests in motion planning algorithms, graphics, robotics, and computational geometry, as well as application areas such as those listed above. Students from Computer Science as well as other departments are welcome. For students in Computer Science, the course project report could serve as a basis for the MS Program Product requirement and/or the department technical writing requirement.

Credit Hours: Variable (see below).

Grading: For 3 credit hours: 40% course project, 20% presentation, 20% assignments, 20% participation. Each student will select the topic of his/her course project. Each student will also present a topic based on a research paper of his or her choice from the research literature and complete a pair of written/programming assignments. For 1 credit hour: 50% paper presentation, 50% participation.

Prerequisites: Knowledge of undergraduate level calculus, linear algebra, and programming (any language, such as Java, Matlab, C, C++, etc.). Prior coursework in motion planning, robotics, or graphics is not required. Both graduate and undergraduate students are welcome to enroll.

Textbook: There is no textbook for this course. Course notes, in-class handouts, and links to relevant papers will be provided.

Tentative Schedule (subject to change)

Date Topics Details
January 12 What is Motion Planning?
January 14 Path Planning for Point Robots
January 19 Introduction to Configuration Space
January 21 C-Spaces and Sampling-based Motion Planning
January 26 Probabilistic Roadmaps
January 28 Improving PRMs
February 2 Single-Query Motion Planning and RRTs
February 4 Bug Algorithms and Introduction to Uncertainty
February 9 Markov Decision Processes Assignment 1 posted
February 11 Markov Decision Processes Continued First paper selection due (category 1)
February 16 Stochastic Motion Roadmaps
February 18 Guest lecture - Camera networks for tracking and navigation
February 23 Guest lecture - Multi-robot Motion Planning I
February 25 Category 1 presentations:
Jinghe Zhang: Rapid Replanning in Dynamic Environments
Lisa Lyons: Toward optimal configuration space sampling
March 2 Category 1 presentations:
Gu Ye: Belief Roadmaps
Eric Harmon: Planning Motion in Completely Deformable Environments
Assignment 1 due
Project proposal due
March 4 Category 1 presentations:
Stan Gregory: Balancing Exploration and Exploitation
Ardavan Kanani: TangentBug: Range-Sensor-Based Navigation Algorithm
Second paper selection due (category 2)
March 9 Spring Break
March 11 Spring Break
March 16 Category 1 presentations:
Tianren Wang: MDPs for multitarget multisensor tracking
Xiaoyang Wen: Apprenticeship learning for navigation
March 18 Category 1 presentations:
Yu Zheng: Multi-modal planning for non-expansive configuration spaces
Pavel Chtcheprov: Probabilistic Cell Decomposition
First project progress report due
March 23 Guest lecture - Multi-robot Motion Planning II
March 25 Special lecture by Ruzena Bacjsy
March 30 Generating Motion
April 1 Machine learning for robots:
Xiaoyang Wen: Quadruped locomotion
Tianren Wang: Robot tracking
Assignment 2 posted
April 6 Uncertainty and Multi-agent systems:
Ardavan Kanani: Bounded uncertainty roadmaps
Stan Gregory: Multi-robot mapping
Second project progress report due
April 8 Learning and control:
Yu Zheng: Learning for grasping
Pavel Chtcheprov: Human-robot skill transfer
April 13 Information spaces and uncertainty:
Gu Ye: LQR-Trees
Eric Harmon: Information spaces for mobile robots
April 15 Flexible objects:
Jinghe Zhang: Steerable electrode arrays
Lisa Lyons: Deformable linear objects
Assignment 2 due
April 20 Final project presentations:
Ardavan Kanani
Gu Ye
Third project progress report due
April 22 Final project presentations:
Yu Zheng
Eric Harmon
Tianren Wang
April 27 Final project presentations:
Lisa Lyons
Katelyn Millay & David Skwerer
Jinghe Zhang
Xiaoyang Wen
April 29 No class Final project report due by noon.
Project report format must be based on IEEE Template for all Transactions (except IEEE Transactions on Magnetics/Photonics).

Suggested Papers for Category 1: The Basic Motion Planning Problem and Uncertainty

Suggested Papers for Category 2: Other Topics in Motion Planning

Relevant Textbooks

There is no textbook for this course. Below are standard textbooks in the field that you may find useful or interesting.

Advice on Presentations

Giving an Academic Talk by Jonathan Shewchuk