Fuzzy relational architecture in robot control based on visual feedback

Fuzzy relational model describes system´s behavior by a nonlinear mapping between its variables and it could be applied for the prediction of system behavior. In this paper, we further develop the ideas of using fuzzy relations in construction of internal models for control purposes. The architecture based on the fuzzy relational model was developed. Fuzzy relations were used to describe the behavior of simple reactive agents. Their task was to predict the existing and desired system´s states and to control the system in desired direction. The procedure was illustrated by example of visually controlled robot system, giving him the ability (property) to reach the desired target point. That was one of the basic building blocks in emergent behavior/functionality of the system.