Mathilda is a whole sujbect by herself. Here will be explained the origin of the project and what we did.
But first, please enjoy a video of Mathilda walking.
This project is now more than three years old and I learned a lot. Everything started when I pointed out to one of my friend that yes, physics is nice, we are learning a lot of things at EPFL. But all we are doing is solving equations on a black board more or less. This is obviously an exageration but you got the point.
So, what kind of really exciting project we could do?
By bringing more friends and ending up with six crazy people, Mathilda was born.
I develop my skills all along this project, mainly on the DIY and manual work side. I learned how to do stick soldering for the paws and I used a lot the lathe for the pipe preparation.
We wanted Mathilda to be approximatively the size of a juvenile T-Rex. It is already a chalenging size, especially for the weight consideration. The first time we made her walk, the walls were vibrating because of the impact!
One of the aim was to have Mathilda to be able to evolve and interact with her environment with the least human control as possible. She could be reacting to external stimuli like a person moving in front of her or to touch.
Mathilda is supposed to be able to move on flattish surface without falling. In order to have a convincing walk, we decided to have a fully mechanical movement, which means she is not self balacing herself using electronic input.
We used mainly refurbished materials for cost reduction. We chose aluminium pipes for the main structure and poly-carbon for the skull and the ribs.
For the motion we only use an electric screwdriver rotating motor and a car battery. Mathilda does not require more to be able to walk. She can be also pretty fast, to give you an idea, on the video she is at the slowest pace.
We are currently working on the way she is walking. We want her to be as close as possible to how a real T-Rex would walk. We studied the skeleton of some fossils and we realized that the hips were too wide.
We will integrate a sound system along with basic vision so she could roar when stimulated by optical events such as a close person moving. We are still not sure on how she should sound like but we are working on a panel of noise she could produce depending on the behaviour we want to express.
For the moment she looks like a bare skeleton. We wanted to add a skin but we had an issue with the heat evacuation and the elasticity. Since recent studies suggested that actually dinosaurs could have been covered with feathers, we thought it would be an awesome solution. The feathers would idealy be able to raise depending on Mathilda's mood of the day.
The perfect situation would be to implement a kind of partern recognition, like a QR code, that she could identify and follow. Just imagine wearing a T-shirt with this QR code and having Mathilda following you on the street!