The knowledge engineering approach to autonomous robotics

Abstract

Given that a robot is no more than a Knowledge-Based-System taking control of the interaction with the environment of a physical platform with sensors and effectors, it seems natural to integrate the tasks and methods used in Autonomous Robotic (AR) in the broad field of Knowledge Engineering (KE). This is the global objective of this paper. We illustrate the approach from a situated perspective. That is, for a specific class of robots in a specific class of environments. The usual methodology in KE is applied to model, operationalize and implement the inferential schemes corresponding to the endogenous modeling of the environment and the navigation tasks. The paper ends with the contributions to a robotic ontology, emerged in the modeling phase. © Springer-Verlag Berlin Heidelberg 2003.