Weblog on DTU RoboCup project - spring 2003

Real-time image processing of an iPAQ based robot with fuzzy logic (fuzzy)

Jesper Damkjær, Jesper Hansson, Steffen Larsen, Steffen Nissen & Sidsel Jensen
damkjaer@diku.dk, hansson@diku.dk, zool@diku.dk, lukesky@diku.dk & purps@diku.dk

University of Copenhagen, Department of Computer Science

The equipment we are using for the project are the following :

1 HP iPAQ 3870 with bluetooth and Familiar Linux
1 Double PC Card Expansion Bay
1 Serial Cable
1 Pontech PIC Controller
1 Lynxmotion SUMO Robot Terminator kit
1 Winnov VideumCam Traveler

The robot has 4 motors, each serial connected to 1 speedcontroller. The speedcontroller is connected to a small PIC, which is connected to the serialport of the iPAQ. The iPAQ controls the camera through a pccard interface in the sleeve. The robot will be autonomous, running on camera input based on a white tape line on the floor.

As a spinoff to this project, we will be making a small application using bluetooth, that can remote control the robot from a Nokia cell phone.

Pictures of the robot so far.

We signed up for DTU RoboCup 2003 competition.

We took a trip to DTU, to check out their robot-building facilities, and also to test our own robot.

See the pictures here

We worked on the neural network, that our robot will use to identify where the line is.
We have made a small qt based gui, that will enable us to train the network more easily, using pictures taken from the robot camera.

On the picture below you can see the gui. The first picture is the normal picture. The second picture is normalized, gaussed and the ridgedetection algorithm has been run (black dots indicate a ridge). The third and last picture indicate the output from the neural network. The color of the boxes indicates how strongly it suggests where the line might be located. We can position the red line where the line actually is, and thereby training the network.