Cooperative spectrum sensing in decentralized cognitive radio system

Cooperative spectrum sensing (CSS) is capable to improve the performance of signal detection. It is one solution to overcome fading, shadowing, and uncertainty noises. Unlike centralized networks, Secondary User (SU) in decentralized networks independently senses license channel without forwarding the sensing result to central coordinator. This paper investigates the performance of Decentralized Multiuser CSS (DMCSS) in Cognitive Radio (CS) system. Partially observable Markov decision process (POMDP) is used as an aid for SU to sense and access a license channel. It models spectrum as Markov discrete process where idle and busy states are represented by '0' and '1'. The comparisons are made against Centralized Multiuser Cooperative Spectrum Sensing (CMCSS). The derived results show that CMCSS has better performance than DMCSS model. However, at certain values of SNR, DMCSS performance outperforms the achievable throughput by CMCSS through setting of collision threshold (ζ). © 2013 IEEE.