A Virtual Receptor in a Robot Control Framework

We propose a general-purpose virtual receptor for 3D robot vision based on RGB-D sensor data. The application independent robot vision framework performs two basic tasks: it creates a 3D metric map of the environment and it recognizes basic 3D solids and 2D textures and shapes. The design methodology follows the principle of knowledge-based systems, as the virtual receptor is structured into a knowledge base (including the model, data and inference rules) and a control strategy. Procedural semantic networks are chosen as the knowledge representation language. Their main features - an object-oriented modeling of the environment and non-monotonic logic of inferences - makes them specially suitable for 3D object recognition in RGB-D images. The interfaces to other modules of a autonomous robot control structure are discussed also - these are: the main control and ontology modules.