*Lijin - another meeting*

Just a few quick notes from my last meeting with Lijin regarding my project. I am having to respecify my project notes so that they are in more technical/scientific terms for all the researchers here. It is also helping me prepare for an upcoming doctoral presentation I will be making for the ‘Selected Topics in Artificial Intelligence‘ series

girltron is:

Under actuated - more DOF’s than motors
Uses novel materials - wood
Has body intelligence

One problem I have found is that there is a large gap between robot research prototypes and commercial robots - and girltron is somewhere in the middle of these two extremes. One of my aims with the project is to produce a robot that is a finely crafted objected, rather than a mass produced item.

She’s also going to be operating in a complex environment in real time and with the general public. Most research prototypes are only operated for limited amounts of time, in very controlled situations. So this is a major factor in designing the hardware and software, I need to make sure she will not break down or malfunction after long periods of use.

Still unsure if I will model her behavior via a state machine or neural network?.. So its been a good thing I have been attending Rolf’s lectures on Neural Nets. If I go with a NN I like the idea of modeling it on the endocrine system, to represent a natural regulating internal system.

Lijin is designing me a distributed controller for the robot, with modular sensors. The controller will also be used with the swimming humanoid she is working on. She’s just organised a Masters student, Marcelo Zanetti to come in and help with this.