On optimal fuzzy c-means clustering for energy efficient cooperative spectrum sensing in cognitive radio networks

This work explores the scope of Fuzzy C-Means (FCM) clustering on energy detection based cooperative spectrum sensing (CSS) in single primary user (PU) cognitive radio network (CRN). PU signal energy sensed at secondary users (SUs) is forwarded to the fusion center (FC). Two different combining schemes, namely selection combining (SC) and optimal gain combining are performed at FC to address the sensing reliability problem on two different optimization frameworks. In the first work, optimal cluster center points are searched for using differential evolution (DE) algorithm to maximize the probability of detection under the constraint of meeting the probability of false alarm below a predefined threshold. Simulation results highlight the improved sensing reliability compared to the existing works. In the second one, the problem is extended to the energy efficient design of CRN. The SUs act here as amplify-and-forward (AF) relays and PU energy content is measured at the FC over the combined signal from all the SUs. The objective is to minimize the average energy consumption of all SUs while maintaining the predefined sensing constraints. Optimal FCM clustering using DE determines the optimal SU amplifying gain and the optimal number of PU samples. Simulation results shed a light on the performance gain of the proposed approach compared to the existing energy efficient CSS schemes.