Autonomous Decision Making Process for the Dynamic Reconfiguration of Cognitive Radios

The aim of this paper is to present a generic cognitive framework for autonomous decision making with regard to multiple, possibly conflicting, operational objectives in the face of a time-varying environment. The framework is based on the definition of two scales introducing order relationships between the configurations that help the reasoning and learning processes. The resulting cognitive engine learns from scratch to identify the optimal configuration with regard to the design objectives given the current radio environment. The proposed approach is validated on a case of cognitive waveform design voluntarily kept simple for illustration purpose. The possibility of extending this framework to more complex problems is also discussed.