Elements of Hybrid Control in Autonomous Systems and Cognitics

Intelligent systems and robots have made significant progress in recent years, in artificial intelligence (AI), situated automata, reactive systems, classical control, computer infrastructure and networks. The paper shows that hybrid approaches (i.e. a mix of these mentioned techniques), and in general cognitics (i.e. automated cognition), offer multiple benefits: assessing various cognitive elements, predicting phenomena, compensating for disturbances, embedding programmed systems in reality, estimating entities in virtually all regions of high dimensional spaces, learning and being expert, and running on digital processors with the best software methods. Additionally, some concepts are newly discussed, in order to facilitate a coherent global view: deliberation, top-down approaches, creativity and ingenuity are also important items in the general picture. The paper finally also addresses learning; it appears that only chance has the cognitive power to yield truly novel models, while expert resources remain necessary to collect, optimize, and make use of them effectively.