An Intelligent Scheme for First Run Cognitive Radios

Cognitive Radio (CR) can access the spectrum temporarily to solve the problem of the near spectrum crunch. The previous transmissions´ events are one of the main motivations for the CR actions and its learning procedures. Therefore, self aware CR devices may cause a considerable interference when they transmit for the first time with no practical knowledge. This paper proposes a solution for cognitive radios with no experiences for instance first run cognitive radios and CR devices repositioned in new wireless environments. A multi-layered learning system is designed using the neural networks to extract cognition from other operator secondary users. Thus, sensors data, other CR devices´ data, spectrum governing entities behaviour, in addition to any stored data in the CR are processed for decision comparisons. Final evaluations are based on the weight and significance given to each of these inputs. As a result, full autonomous mature cognitive users are created through understanding other cognitive radios experiments. Simulation was trained to assess the proposed learning model. Results show promising and efficient utilization using gradual learning rate algorithms with the designed system.