[Adaptive Robot Controller]

Date: May 2011

A robot neural controller where obstacle-avoidance and random-walk neurons observe level of their irritation and randomly reconfigure their connections to motors when the irritation reaches a specified threshold is presented and tested.

Individual neurons have self-recurrent connections and receive function-specific input. The neural controller shows either robustness or adaptivity to noise and sensory and motor function perturbations by utilising ultrastability and subsumption principles.

Keywords:
Obstacle avoidance, random walk, artificial neural network, ultrastability, homeostasis
DOWNLOAD PAPERRUN JAVA APPLETDOWNLOAD CODE


{Please enable JavaScript in order to post comments}

Ultrastable Neuroendocrine Robot is born

 I have finally put a beginning of an artificial brain to my Ultrastable Neuroendocrine Robot - UNER... my lego robot can now measure inputs from its 3 front proximity sensors and individual cells in its brain get activated when it sees something

Does Communication Make a Difference?

 This paper compares different animal groups from eusocial insect colonies to human society and discusses their mechanics and behaviour as agent systems. The main focus is on interaction between the agents and on how properties of a system like effectiveness or predictability are affected by these interactions.

Robot Flocking: Sensors and Control

 This paper discusses various kinds of robot sensory input, approaches to motor control and ways they could be used for flocking. Focus is put on vision and Gibsonian optic flow that could be utilised by robots with advanced behaviour.

pyCreeper

 The main purpose of pyCreeper is to wrap tens of lines of python code, required to produce graphs that look good for a publication, into functions. It takes away your need to understand various quirks of matplotlib and gives you back ready-to-use and well-documented code.

Novelty detection with robots using the Grow-When-Required Neural Network

 The Grow-When-Required Neural Network implementation in simulated robot experiments using the ARGoS robot simulator.

Fast Data Analysis Using C++ and Python

 C++ code that processes data and makes it available to Python, significantly improving the execution speed.