Browsing articles in "Projects"

Dolly Tripod Platform

Jan 27, 2011   //   by Martijn   //   Projects  //  Comments Off

In September I found a small collapsable trolley. When I bought it I knew I wanted to use it to make smooth camera moves like a dolly.

The trolley is strong, yet light weight. The only real problem is that it is only 40cm wide and about 70cm long. This means that my tripod could not be placed on top of it while fully extended. In addition to this problem, the trolley doesn´t have a flat surface. It has four metal extendable beams. I concluded that a structure had to be built on top of it.

After considering various materials I decided to make it out of wood. It’s a cheap material and I have some experience with manipulating wood. Since the platform needed to be 80cm wide I got a plank that was 20cm by 160cm.

I attached small wooden blocks on the underside of the platform so that it would fit snugly over the beams of the trolley.  A series of large holes in the top of the platform allow the tripod to stand steadily in fixed positions.

The following video is some test footage made with the dolly.

LarCT-R: Semantic Web for (j)ACT-R

Dec 5, 2010   //   by Martijn   //   Projects  //  Comments Off

As a follow-up project to my master thesis I developed an Eclipse plug-in called LarCT-R.

The plug-in is an extension of jACT-R that allows ACT-R models to retrieve RDF data and regulate LarKC workflows.
The plug-in has been released as an open-source project on

Robot Crawler

Mar 3, 2010   //   by Martijn   //   Projects  //  Comments Off

Recently a friend of mine, Michiel Bouman showed me a project he was working on. It was a robot arm on a cart that could potentially walk around the room. I had a bit of time and offered to come over to work on this project.

First I refactored his programming code to allow for a higher level of abstraction. After that we set out to make the robot teach itself to walk using Machine Learning techniques.  We looked at well know techniques like neural networks for which there were already some useful libraries. However in the end we were researching more specific task learning techniques like Q-Learning.

Unfortunately the project has been shelfed for later continuation, but it would be cool to pick it up again.