Emergence, Artificial Neural Network, Robotic System, Robot Control.

The paper deals with the emergence of behaviour of an autonomous robot controlled by a Back-Propagation neural network. The theoretical part focuses on robotic architecture, and on the issues of emergence and connectionist networks. The output of the experimental study is a 2D application which allows to model the movements of the robot in a simulation environment with the possibility of configuration variability. In addition, several robot type designs are proposed and tested in the experimental part of the article which address the role of movement in an environment aimed at avoiding obstacles.