. . CSE398-011/498-010 Robocup MWF 11:10-12:00 PM Spring 2005

Announcements

• 03 May 05:  Here are a couple of sample programming questions for the final.  Again, I will not post solutions but if you have questions please stop by. 
  
• 02 May 05:  The course wrap lecture with the final overview. 
  
• 02 May 05:  Here are some sample final questions to give you an idea what I'll be looking for.  I will add a programming question or two later today or tomorrow.   
  
• 05 Apr 05:  Team Challenge 2 results.  Every dog deserves its day in the sun.  Apologies for the delay in posting.   A "*" indicates disqualification for the round.  Time in green indicates successful ball detection.  Time in red indicates the dog completed the traial without a collision.  Trial 2 was declated a traw to the tolerances of the timer.  RR's submission for the last trial can be watched here. 
  

	Color Segmentation	Trial 1	Trial 2	Trial 3	Trial 4	Trial 5
A-Team	2.5 m	*	120	*	*	*
Awesome	2.0 m	120	120	*	*	69
MacGyver	2.0 m	120	60.78	120	*	120
Reading Rainbow	3.0 m	117	60.13	*	*	83

• 05 Apr 05:  Team Challenge 3 results (trial winner in bold, green indicates successful trial, red indicates the trial failed to meet the localization distance and time criteria, and * indicates that the distance was measured to the center of the target).  NOTE:  The results have been updated to better reflect my sometimes cryptic notes. 
  

	Trial 1	Trial 2	Trial 3	Trial 4	Trial 5	Trial 6
A-Team	51	54	70	*	21	30
Awesome	*	*	*	*	*	19
MacGyver	17*	*	*	37	21	5*
Reading Rainbow	*	*	*	*	*	*

• 03 Mar 05:  I've added a link to Alonzo Kelly's excellent lecture notes.  The undertainty notes will be of great use when we discuss position estimation and filtering approaches to robot localiaztion. 
  
• 27 Feb 05:  Homework 1 solutions are here. 
  
• 24 Feb 05:  Team McGyver's winning entry into Team Challenge 1 (divx codec required).
  
• 23 Feb 05:  Here is a link to a Robocup/Tekkotsu Seminar given at the University at Albany.   The state machine and vision tutorials by Shawn Turner are especially useful. 
  
• 21 Feb 05:  Team Challenge 2 is posted.
  
• 21 Feb 05:  Prior to the start of class on Wednesday, each team member should make the team report and ALL source code available online.  These URLs should be e-mailed to the entire class.  If you do not have a location to post to, contact me and we will link off of the course web page. 
  
• 21 Feb 05:  Team reports are due at the start of class on Wednesday.  No late reports will be accepted.
  
• 21 Feb 05:  Team Challenge 1 results (trial winner in bold, red indicates trial ended in collision, green indicatates trial ended with successful exit).
  

	Left Trial (s)	Center Trial	Right Trial
A-Team	110	115	55
Awesome	50	45	40
MacGyver	28	75	69
Reading Rainbow	52	32	38

• 16 Feb 05:  Representatives from each team are required to be at the lab (PL450) this Monday 11AM for the first Team Challenge.
  
• 16 Feb 05:  First team presentations are rescheduled for 23 Feb 05. 
  
• 16 Feb 05:  I will be on travel this Friday.  Class is rescheduled for Wednesday, 23 Feb 05.
  
• 10 Feb 05:  The course grading distribution has been changed to reflect the Team Challenge segments. 
  
• 09 Feb 05:  The first Team Challenge is posted. Time to get those competitive juices flowing!
• 09 Feb 05:  Labs are back, thanks to some awesome Linux technical support. We will be working on Fedora from this point on.
• 26 Jan 05:  I've added a link to "Planning Algorthms" text by Steve Lavalle.  You can download the text for free here.  There is also a link to his notes on motion planning. 
• 25 Jan 05:  I've added a link to a graph theory tutorial from the University of Saskatchewan.  It has a gentle introduction to graph theory, and then details on a few algorithms (to include Dijkstra's).  Recommend reading for all students.  
• 17 Jan 05:  First day of classes.  Welcome!

Course Description
This course is inspired by the 4-legged league International Robocup Competition, where teams of Sony Aibo robots are programmed with autonomous behaviors and play competitive soccer (aka football).   In this course, we will use the Aibo platform and the Robocup competition as a medium for inderstanding and applying relevant algorithms from mobile robotics and computer vision.  Students will be paritioned into 4 teams.  The end goal of the course is for the teams to compete in the Lehigh Open 4-legged Robocup Championship.

Topics that will be covered during this course will include (among others):

• Motion Planning
• Computer Vision
• Sensor Fusion
• AI

Prerequisites

• Math 23 Analytic Geometry and Calculus III or equivalent
• Math 205 Linear Methods or equivalent or equivalent
• CSE 109 Systems Programming or equivalent programming experience in C/C++

Instructor
    John Spletzer
    Office Hours:  MWF 10:00-11:00AM and by appointment

Teaching Assistant
    Xiaoguang Qi
    Office Hours:  TBD

Course Location
    Lectures:  MG110
    Laboratory Sessions:  PL450

Lecture Times

• MWF 11:10-12:00PM initially.  MW lectures will be eliminated (for the most part) once the lab sessions begin. 
  
• Regularly scheduled class times will be used for formal lectures, team presentations, weekly team meetings with the professor, as well as laboratory time.
• In addition to the regularly scheduled class times, students will be required to schedule 4 hours/week of laboratory time for working on the Aibo platforms.  This is the MINIMUM required lab time, and will serve in lieu of homework assignments for the course.  Additional time can be scheduled with the instructor on a case by case basis.

.
Course Materials
.
We will rely primarily on the class lecture notes, augmented by the following online materials.  There is no required text for this course. 

1. The official Robocup site can be found here.
2. Our interest is in the 4-Legged League competition.
3. The league rulebook can be found here.
4. All code will be implemented in the Tekkotsu framework developed by CMU.  The beginner's tutorial is required reading. 
5. This Open-R SDK for the Aibos is provided by Sony.
6. Since the lab computers are primarily Windows boxes, we will be relying upon the Cygwin emulator in this course.
7. The Aibo's vision module is based upon the CMVision software described in the paper by James Bruce et al. 
8. Here is a graph theory primer from the University of Saskatchewan.
9. A. Kelly, Introduction to Mobile Robots Course Notes, Field Robotics Center, Carnegie Mellon University, 2000, with a particular emphasis on "Position Estimation 2" and "Uncertainty 1-3"

Grading (Tentative)
.
    Class Participation:  25%
    Team Challenges and Reports:  25%

    Team Presentations and Final Report:  25%
    Final Examination:  25%

• The majority of your course grade will be based upon class participation.  Students are expected to make ALL required lab hours.  Remember, this is a TEAM exercise.  When you are absent, you hurt your whole team.
• Each team is required to meet weekly for approximately 30 minutes with the instructor to discuss progress. 
• There will be approximately 3 team presentations over the course of the semester.  These will be made to the entire class.  Students are expected to discuss accomplishments to that point, plans over the next several weeks, success stories, lessons learned, etc.
• A final report is required that summarizes the  
• In addition to traditional responsibilities, graduate students sitting in on the course are expected to assume a leadership role.  Each will be designated as team leader, and be repsopnsible 
• Grades will be assigned acording to the following distribution:

.

A	92-100	C	72-76
A-	90-91	C-	70-71
B+	87-89	D+	67-69
B	82-86	D	62-66
B-	80-81	D-	60-61
C+	77-79	F	00-59

University Policy on Disabilities:  If you have a disability for which you are or may be requesting accommodations, please contact your professor and the Office of Academic Services, Room 212, University Center or call (610-758-4152) as early as possible in the semester.  University policy states that you must notify your professor seven (7) days prior to the exam.


Tentative Schedule
.

DATE	TOPICS	NOTES
17 Jan 05	Course overview	Lecture 1
19 Jan 05	CMVision & Color Segmentation	Lecture 2 CMVision Paper
21 Jan 05	Edge Segmentation	Lecture 3
24-26 Jan 05	Motion Planning Part 1	Lecture 4 Lecture 5 Graph Theory Primer R. Beard's Paper on Potential Fields S. Lavalle's Motion Planning Notes
28 Jan 05	Intro to the Tekkotsu Framework	Lecture 6 pdf ppt Tekkotsu Tutorial
31 Jan - 2 Feb 05	Laboratory Session 1	Tekkotsu Tutorial
04 Feb 05	Coordinate Transformations	Lecture 7 pdf ppt
07-09 Feb 05	Laboratory Session 2	Team Challenge 1
11 Feb 05	Sensor Fusion Part 1	Lecture 8 pdf ppt
14-16 Feb 05	Laboratory Session 3
21 Feb 05	Team Challenge 1 at PL450
23 Feb 05	Team Presentations I  - New Date!
21-23 Feb 05	Laboratory Session 4	Team Challenge 2
25 Feb 05	Robocup Paper Review 1:  CMU	Lecture 9 pdf ppt
28 Feb 05	Team Challenge 2 at PL450
28 Feb - 02 Mar 05	Laboratory Session 5
4 Mar 05	Localization Part 1	Lecture 10 pdf ppt
08-12 Mar 05	Spring Break
14-16 Mar 05	Laboratory Session 6	Team Challenge 3
18 Mar 05	Localization Part 2	Lecture 11 pdf ppt
21-23 Mar 05	Laboratory Session 7
25 Mar 05	Localization Part 3	Lecture 12 pdf ppt Welch & Bishop's Kalman Filter Tutorial
28-30 Mar 05	Laboratory Session 8
30 Mar 05	Team Challenge 3 - Localization
01 Apr 05	Team Presentations II
04-06 Apr 05	Laboratory Session 9	Team Challenge 4
08 Apr 05	The Extended Kalman Filter	Lecture 13  pdf ppt
12-14 Apr 05	Laboratory Session 10
16 Apr 05	Localization Finale - EKF/PF	Lecture 14  pdf ppt
19-21 Apr 05	Laboratory Session 11
22 Apr 05	Lecture Cancelled
25-29 Apr 05	Lehigh Open Tournament and Wrap Up