Neural Network Data Fusion for Cognitive Radio Network

Abstract

Cognitive radio network practices dynamic spectrum access for efficient spectrum utilization. Sensing being the utmost task involves cooperation of different secondary users contributing to cooperative network. Thus, fusion rules play an important role for decision making. In this paper, we considered a scenario of cooperative spectrum sensing and analyzed the performance of fusion schemes implanted at fusion center for decision making. We have also proposed a learning-based neural network fusion scheme for reliable decision. Further, we have analyzed the performance of proposed scheme under different fading channels. The proposed scheme outperforms conventional schemes accomplishing substantial enhancement in detection probability and decline in probability of false alarm.