Feeling Artificial Intelligence for AI-Enabled Autonomous Systems

The need for more advanced Unmanned Systems (US) is supported by the development trends of the world society. AI plays an important role in supporting the required level of autonomy. AI-enabled US developers focus on Feeling AI (FAI). The considered approaches of cognitive sciences and AI do not fully support such US criterion as autonomy. This article proposes FAI based on the conceptual model of L. Zadeh “Computing With Words” (CWW) to support US autonomy. Proposed FAI architecture bridges the gap between two paradigms “sensor data” and “word”, which is the main challenge for deployment the CWW in US. Blueprint FAI architecture, which satisfies the requirements of autonomy, implemented on the basis of cognitive models and is presented as a set of homogeneous knowledge granules, is considered. Architecture includes four knowledge bases and supports nine engines, including Perception, Short Term memory, Control, Goal-Driving, Planning, Attention, Emotion and Need Engines. Homogeneity is achieved due to the unification of the knowledge granular in the form of external and internal meanings of the word. Formal representation models and processing algorithms based on fuzzy systems are given. A numerical example is considered.