CS 498: Introduction to Planning Algorithms
Fall 2011
Tue/Thu 12:30-1:45
Room 4407 Siebel Center
Registration: 40091 (3 hrs), 40092 (4 hrs)
Instructor:
Steve LaValle
Office Hours: Tue/Thu 2:00-3:00
A Google translation
from Chinese to English of
the textbook summary:
Planning is the crystallization of human wisdom,
planning issues widely found in people's daily work and life. Now,
the planning has been involved in computer science, artificial
intelligence, mechanical, mechanics, control theory, game theory,
probability theory, graph theory, topology, differential geometry,
algebraic geometry and many other modern sciences. "Planning
Algorithms" is the author's years of teaching and research summary, a
systematic introduction to the basics of planning areas and the latest
results. The author of three independent disciplines: robotics,
artificial intelligence and cybernetics cleverly combined. "Planning
algorithm" gives a lot of informative examples to have been relatively
difficult to understand the mathematical problem becomes to life,
after-school reading references and exercises is to further deepen and
expand the reader's understanding of appropriate content.
Course Calendar (tentative):
Class #
| Date
| Lecture Topic
| Reading
| Assignments
|
Week 1
|
1 | 8/23 | course overview
| Ch. 1 |
|
2 | 8/25 | discrete feasible planning; search algorithms
| Sec. 2.1, 2.2 |
|
Week 2
|
3 | 8/30 | geometric representations; transformations
| Sec. 3.1,3.2 |
|
4 | 9/1 | transformations, topological concepts
| Sec. 3.2,4.1 |
|
Week 3
|
5 | 9/6 | topological concepts
| Sec. 4.1.1 | HW1 Assigned
|
6 | 9/8 | manifolds
| Sec. 4.1.2 |
|
Week 4
|
7 | 9/13 | manifolds, C-space
| Sec. 4.2,4.3.1 |
|
8 | 9/15 | C-space, C-space obstacles
| Sec. 4.3.2 | HW1 Due
|
Week 5
|
9 | 9/20 | C-space obstacles
| Sec. 4.3.2 |
|
10 | 9/22 | metric spaces
| Sec. 5.1.1, 5.1.2 |
|
Week 6
|
11 | 9/27 | sampling theory, collision detection
| Sec. 5.2, 5.3 | HW 2 Assigned
|
12 | 9/29 | collision detection, incremental sampling and searching
| Sec. 5.3, 5.4 |
|
Week 7
|
13 | 10/4 | incremental sampling and searching, RRTs
| Sec. 5.4, 5.5 |
|
14 | 10/6 | RRTs, roadmap methods
| Sec. 5.6 |
|
Week 8
|
15 | 10/11 | combinatorial motion planning, polygonal obstacles
| Sec. 6.1, 6.2 | HW 2 Due
|
16 | 10/13 | Midterm Exam
| . |
|
Week 9
|
17 | 10/18 | CLASS CANCELLED
| . | HW 3 Assigned
|
18 | 10/20 | vector fields, differential constraints
| Sec 8.3.1, 13.1 |
|
Week 10
|
19 | 10/25 | differential constraints
| Sec. 13.1 | .
|
20 | 10/27 | car-like robots;
differential drive
| Sec. 13.1 |
|
Week 11
|
21 | 11/1 | phase space, double
integrator, particle motions
| Sec. 13.2.1, 13.3.2 | .
|
22 | 11/3 | motion planning with differential constraints
| Sec. 14.1 |
|
Week 12
|
23 | 11/8 | sampling the control
space; incremental sampling and searching
| Sec. 14.2, 14.3 | HW 3 Due
|
24 | 11/10 | incremental sampling and searching
| Sec. 14.3,14.4 |
|
Week 13
|
25 | 11/15 | physical sensors, virtual sensors
| Sec. 2 and 3 of this | HW 4
Assigned
|
26 | 11/17 | virtual sensor models
| Sec. 3 of this |
|
Thanksgiving Break
|
Week 14
|
27 | 11/29 | more virtual sensors; sensor lattice
| Sec. 3 of this |
|
28 | 12/1 | spatial and temoporal filtering
| Sec. 4 of this |
|
Week 15
|
29 | 12/6 | planning in information spaces
| Sec. 5 of this | HW 4 Due
|
Homework Assignments
- HW1: Assigned: Sep. 6, Due: Sep. 15.
Solve the following problem: Using the Forward Search Template
for discrete feasible planning, suppose that Q is sorted so that
the element that is waiting on the queue for the median amount of
time (the middle element in terms of time waiting) is selected by
GetFirst(). Is this method systematic for the case of a finite
search graph? Is it systematic for a countably infinite search
graph? Explain your answer.
Also, solve book problems 3.2, 3.14. 4.4, (4.17: you can instead turn
in this problem as part of HW 2). Problem 3.14 involves an
implementation. You may use any language you like. I recommend using
python and the PyGame package (it is easy to install on most
platforms, including Ubuntu). Here is some sample
code: [1][2]. You may
assume that each link is a rectangle. Allow any number of links to be
transformed, as a parameter in your code. Display the results for
various examples.
Turn in written (or printed) copies of your work in class.
- HW2: Assigned: Sep. 25, Due: Oct. 11.
Solve the
following problems from the book: 4.8, 4.17 (implement the algorithm
and show the results for various cases), 4.18, 4.19, 5.2, 5.6.
- HW3: Assigned: Oct. 18, Due: Nov. 8.
Extend your
implementation from HW 1 to implement a bidirectional RRT algorithm.
Allow the robot to be a chain of line segments, for which the first
link in the chain is attached to a fixed position (but the link can
rotate about that point). Allow the obstacle region to be any
collection of discs of various radii. The discs may overlap. One
component of the assignment is to develop a simple collision detection
module that determines whether the robot is touching any of the discs.
Also, assume each link of the robot is not allowed to touch other links, except
the consecutive links, to which it is attached by a joint. The
configuration space is a d-dimensional torus, for a d-link robot. The
user should be able to specify a d-link robot, a collection of
obstacles, an initial configuration and a goal configuration. Your
software should determine a collision-free path and animate the
solution. You must design some interesting examples to demonstrate
the algorithm. Turn in your code and some printouts of the solutions.
- HW4: Assigned: Nov. 17, Due: Dec. 6.
Solve the
following book problems: 6.2, 6.4, 13.1, 13.2, 14.2(a)
Topics
(estimated number of lectures given):
- Introduction, motivation (1)
Chapter 1
- Discrete planning (2)
Sections 2.1, 2.2
- Continuous planning, sampling-based methods (8)
Sections 3.2, 4.1, 4.2, 4.3.1, 5.1-5.4, and possibly 7.1, 7.2, 7.6
- Motion planning under differential constraints (6)
Sections 8.3, 13.1, 13.2, 14.1-14.3
- Planning under uncertainty in predictability (7)
Sections 2.3, 9.2-9.4, 10.1-10.4
- Sensors and information spaces (general concepts, sensor models,
minimal information requirements, examples) (6)
Tutorial paper and parts of Chapters 11 and 12.
Textbook:
Readings above are from Planning Algorithms, S. M. LaValle,
Cambridge University Press, 2006. Also available for free download at
http://planning.cs.uiuc.edu/.
Course Mechanics
- Midterm (30%), Final Exam (40%), Homeworks (30%)
- An independent project is required for those registered for four
credit hours.
- Homeworks will be a combination of written and programming
assignments. There are no requirements to use a particular
language for programming assignments. There will be three or
four homework assignments during the semester.