Anna Pawlicka

Programmer. Hiker. Cook. Always looking for interesting problems to solve.


About | Archive | Talks

Continuous Time Recurrent Neural Network (CTRNN) for e-puck

26 Feb 2013 | CTRNN, e-puck, robotics, Webots

After performing some experiments with Q-Learning algorithm, reading tons of research papers and finally realizing that evolving walking in Aibo is a research project on its own, I decided to change the simulation to e-puck. e-puck is a differential wheels robot, kindly nicknamed “yoghurt pot” by my friends:

Why e-puck? Only two wheels to control (as opposed to 16 joints in Aibo). As a bonus, once the algorithm is completed I’ll have a chance to test it on a real robot!

The algorithm I’m implementing is Continuous Time Recurrent Neural Network (CTRNN). It’s just a first step. Second step will be to evolve two populations: agents and games. Games will represent different fitness functions, agents – neural networks. I’m back to evolutionary computation.


Older · View Archive (25)

Q-Learning for Aibo

I’ve been looking more into motor actions and their effect on Servos.  The initial position is always zero. So each time a new action is applied to the motor (we_servo_set_position()), it is interpreted as absolute position. In order to emulate a relative position, I’ll have to store the last value passed to the method, and then just add that value to the newly computed one. Not sure if this will improve my algorithm as it seems that relative position will eventually result in maxing out the joint position.

Newer

Deadlines help you achieve the impossible.

Just 10 days left until I have to submit my draft report. This describes how I feel right now:

Super snail