Cognitive and robotic systems: Speeding up integration and results

Abstract

Nowadays, state-of-the-art robots are capable of millimetric motion accuracy by performing highly repetitive tasks, however, as a constraint they operate in highly structured environments where objects are in known and predictable locations. Thus, it is not surprising that robots are more often used in high-volume operations such as painting and welding, rather than operations where diversity of actions, direct contact with humans, and variability of the environment meet fundamental requirements. Social robotics is an area of research that aims to make viable the direct interaction of robots with humans in unstructured environments. It uses several techniques, such as, machine learning, cognitive modeling, artificial intelligence, knowledge representation and ontology. One factor that compromises the rapid evolution of social robotics is the difficulty in modeling cognitive systems due to the volume and complexity of information produced by a chaotic world full of sensory information. In addition, the validation of results with the use of real environments involving buildings, streets and people presents a high cost of installation and maintenance. This article offers two strategies to speed up the evolution of social robotics. The first involves the definition of OntCog ontology that models the senses captured by the agent robotic sensors. This modeling facilitates the reproduction of experiments associated with cognitive models and the comparison among different implementations. The second is associated with the development of Robot House Simulator (RHS), which provides an environment where a robot and a human character can interact socially with increasing levels of cognitive processing. An unprecedented feature of this simulator is to provide information about all the senses of the robot, actually only the sense of vision or touch has been considered in the existing robotic simulators.