Behaviour-based architecture for abstract learning and control
Pejman Iravani · 2004
This paper presents a modular behaviour-based architecture for the control and learning of behaviours in mobile robots. The main contribution of this paper is the integration of techniques for generating grounded symbols from raw sensorymotor data and traditional behaviour-based architectures. The resulting architecture reduces the intrinsic dimensionality problems of robotics and allows for abstract control. The architecture is experimentally validated on a simulated robot environment. 1.