Institut für Informatik, Universität Freiburg

Praktikum Robotik WS 2002/2003

AG Künstliche Intelligenz

 

Description

This Praktikum provides for motivated students the opportunity to apply their knowledge in the field of RoboCup. This may include the areas of artificial intelligence, image/sensor processing, real time systems and control engineering. For a more detailed description of RoboCup have a look at www.robocup.org.

Our current activities are the soccer playing robots (CS Freiburg) and a robot capable of playing the game table soccer (KIRO). For a detailed description of the soccer robots have a look at www.cs-freiburg.de and for the table soccer robot at www.informatik.uni-freiburg.de/~kiro/english/index.htm.

Students participating in this Praktikum will have to choose from assignments of one of these two projects. In this context, they will learn how robots percept their environment, decide which actions have to be executed and how to execute them reliably.

Requirements

Since all of our software is written in C/C++, programming experience is absolutely necessary. Additionally, it would be useful if students are familiar with techniques from the area of artificial intelligence, image processing or control engineering.

We are particularly interested in students that intend to continue their work within our group. They might also get the opportunity to attend to international competitions. To reach this goal, however, we expect highly motivated students.

Time and Location

Time: Once a week on Thursday from 14.00 - 18.00
Location: Building 82
Preliminary discussion: 17.10.2002 in the RoboCup room, building 82 (Mensa), entrance through the blue door

Organisation

Alexander Kleiner
Email: kleiner@informatik.uni-freiburg.de
Phone: 203-8226
Fax: 203-8222


Thilo Weigel
Email: weigel@informatik.uni-freiburg.de
Phone: 203-8229
Fax: 203-8222

Talks

Assignments

Robot Control

Vision Features




Getting started

In order to get the code for your project running please do the following steps:


    Compilation

    1. Login with your user (any machine in the pool)
    2. Checkout the required software:
            cvs co csfreiburg
            cvs co csfreiburg_libs
    3. Choose your directory:
            Vision project:
                    cd ~/csfreiburg/omnivision/
            Control Project:
                    cd ~/csfreiburg/robot/ or
                    cd ~/csfreiburg/sensors
    4. Make the code:
            make depend
            make
 

     Execution

        1. Copy the executable on the robot of your choice:
                cd ~/csfreiburg/
                make distpraktikum C=the_name_of_your_robot's_laptop
        2. Remote execution:  
            2.1. Allow the laptop to send a window to your machine:
                    xhost +the_name_of_your_robot's_laptop
            2.2. Login on the laptop:
                ssh praktikum@the_name_of_your_robot's_laptop
            2.3. Run your application:
                e.g. omnivisiontest or sensorsTest
        3. Execution on the laptop (better for vision)
             3.1 Kill the running X-server (ctrl-alt-backspace)
             3.2  Login as user praktikum
             3.3  Execute
  

In order to modify the code, please have a look to our simple manual.


Good Luck!




Papers and Documentation


Control

Jack Baltes Real-time Control Theory (PID) Talk, 1999. (download)


Kalman-Filter / Dead Reckoning

J.-S. Gutmann, T. Weigel, and B. Nebel. Fast, accurate, and robust self-localization in polygonal environments. In Submitted, 1999.( download)

Kurt Konolige Robot Notes 2001( download)

G. Welch and G. Bishop. An introduction to the kalman filter. Technical Report TR 95-041, University of North Carolina, Department of Computer Science, 1995.(download)

J.S. Gutmann, Robuste Navigation autonomer mobiler Systeme, Dissertation,^M Fakultät für Angewandte Wissenschaften, Univ. Freiburg, Okt. 2000. (download)

Markov Localization

D. Fox, W. Burgard, and S. Thrun. Markov localization for mobile robots in dynamic environments. Journal of Artificial Intelligence Research, 11:391--427, 1999. (download)

D. Fox, W. Burgard, F. Dellaert, and S. Thrun. Monte Carlo Localization: Efficient position estimation for mobile robots. AAAI-99. (download)


Behavior-Networks

K. Dorer, Motivation, Handlungskontroler und Zielmanagement in autonomen Agenten , Dissertation, Fakultät für Angewandte Wissenschaften, Univ. Freiburg, Feb. 1999. (download)

K. Mueller, Roboterfussball: Multiagenetnsytem CS Freiburg , Diplomarbeit, Fakultät für Angewandte Wissenschaften, Univ. Freiburg, 2000. (download)


RoboCup Teams

T. Weigel, J.S Gutmann, M. Dietl, A. Kleiner, B. NebelCS Freiburg: Coordinating Robots for Successful Soccer Playing , Fakultät für Angewandte Wissenschaften, Univ. Freiburg, 2002. (download)

CS Freiburg Team Description , Fakultät für Angewandte Wissenschaften, Univ. Freiburg, 2002. (download)


Vision

A. Topor, Die Ballerkennung und Positionsschätzung des CS-Freiburg Teams beim RoboCup'98, Studienarbeit, Fakultät für Angewandte Wissenschaften, Univ. Freiburg, Feb. 1999. (download)

J. Bruce, T. Balch, and M. Veloso. Fast and cheap Omage Segmentation for Interactive Robots. 1999. (download)

paper 1 (download)

paper 2 (ownload)

paper 3 (download)

paper 4 (download)


Pathplanning

A. Topor, Roboterfußball: Pfadplanung in dynamischer Umgebung, Diplomarbeit, Fakultät für Angewandte Wissenschaften, Univ. Freiburg, Feb. 1999. (download)


Reinforcement Learning

A. Kleiner and M. Dietl and B. Nebel, Towards a Life-Long Learning Soccer Agent, Proceedings of the International RoboCup Symposium '02, Fukuoka, Japan, 2002. (download)

Documentation

The CS Freiburg Team, Crystal, the interface for the CS Freiburg Robots, Dokumentation, Fakultät für Angewandte Wissenschaften, Univ. Freiburg, Oct. 2001. (download)