Since the last few decades, research in the area of robotics technology has been emphasizing in the modeling and development of cognitive machines. A cognitive machine can have multiple cognitive capabilities to be programmed to make it artificially intelligent. Numerous cognitive modules interact to mimic human behavior in machines and result in such a heavily coupled system that a minor change in logic or hardware may affect a large number of its modules. To address such a problem, several middlewares exist to ease the development of cognitive machines. Although these layers decouple the process of logic building and communication infrastructure of modules, they are language-dependent and have their limitations. A cognitive module developed for one research work cannot be a part of another research work resulting in the re-invention of the wheel. This paper proposes a RESTful technology-based framework that provides language-independent access to low-level control of the iCub’s sensory-motor system. Moreover, the model is flexible enough to provide hybrid communications between cognitive modules running on different platforms and operating systems. Furthermore, a cognitive client is developed to test the proposed model. The experimental analysis performed by creating different scenarios shows the effectiveness of the proposed framework.
Learning in robotics. Learning Goal-directed Behaviour as Dynamical System for Sensory Motor System.
