Robot Eksperimentarium 2008

From ImageWiki

Jump to: navigation, search

This page contains various informations that are relevant to participants of the course "Robot Eksperimentarium", 2008.


Contents

Tips and Tricks: Share your hard earned knowledge


Groups

To keep track of which students are working together, we have this list of groups. Please make sure that you're name is in the right group.


Students following the course

Here is a page showing the students following the course in 2008 (2007).

Exercises

The exercises for the course can be found in Robotopgaver 2008 (previous years: 2007, 2006) (in Danish).


The Robots

This year we'll be using the Evolution Scorpion Robots. For controlling the robot hardware we'll be using ERSP or Player / Stage, and for image processesing we'll be using OpenCV. If you want to use Player you have to use our newly developed Player / Stage driver. The documentation can be found here and notes on further development maybe found here.

To use the robot you'll need access to a computer with ERSP installed. We have 6 such laptops that you need to share. Laptops will be handed out to groups and the group members will have to figure out how to share. We will most likely be using the rule 6 people for 1 laptop or two groups of 3 persons each.

If you're having any problems with the laptops for the robots, please contact Kim Steenstrup Pedersen. You can also report you problems at the bottom of this page (please write your name and the date of report on the page as well).


Tutorials

The SVN Repository

There is a subversion repository for the course. Since you can't be garantied to work on the same laptop twice it is recommended to use the SVN repository to keep track of your code.

To use the repository you need an account on image.diku.dk, which can be acquired by contacting Kim Steenstrup Pedersen.

To work with the code, the first time only you must make a local copy on your local account via the checkout command:

svn checkout svn+ssh://yourLogin@image.diku.dk/home/robot/svn/trunk/ Robot-GROUP_NUMBER/

Where GROUP_NUMBER is the number of your group. You might need to enter your password several times. Please note that you need biology online external access, so from the DIKU UNIX system, you need to work on tyr, ask, etc..

This should create a local directory called Robot-GROUP_NUMBER/ on your local account containing the trunk of the repository. Later revisions can be obtained by,

cd Robot-GROUP_NUMBER/
svn update

Your group code should go into

Robot-GROUP_NUMBER/Robot-GROUP_NUMBER/

To write to the repository simply use

svn commit -m "My short description of what I did for the record"

The work recommended workprocess is:

Before you start the days work do:
 cd Robot-GROUP_NUMBER/
 svn update
Make your editing and once in a while and definitely when your are done do:
 svn update
 svn commit -m "My short description of what I did for the record"

If you want to add a new file, say myfile, you have to do the following after creating and editing myfile

 svn add myfile

If you want to copy example and utility code already in the repository into your group directory, the easiest way to also get it into svn is by the following:

 svn copy src dst

where src is the the source files or directory, and dst is the destination, probably Robot-1/ (if you are in group 1).

If you need access to your files, but don't have a SVN client installed, you can use the web frontend at http://image.diku.dk/svn/robot/.

More information on subversion may be found on freelance resume writers and http://subversion.tigris.org.

ERSP

The robotics platform used on the course is called ERSP (Evolution Robotics Software Platform, I think). You can find a short tutorial on how the platform works here.

Player/Stage

Documentation for Player/Stage can be found here and documentation of the DIKU ERSP Player driver can be found here.

OpenCV

For image processesing we use the OpenCV library. The OpenCV wiki is the place to find help and it includes online reference manuals. We also have a short Danish tutorial on OpenCV.

The Robot Camera

The robots use a Logitech Quickcam Pro 3000 webcam to acquire images. The internal parameters of this camera can be changed since it uses the pwc driver.

The camera has been calibrated to compensate for the Barrel distortion. See Kalibrering af Scorpion robotternes kamera (in Danish).

Object tracking

If you need to track an objects position in an image sequence, you can use the Two Stage Hybrid Tracker.

Personal tools