Wideband spectrum sensing based on energy detection for Cognitive Radio network

Spectrum sensing is an essential concept in Cognitive Radio (CR) systems. It exploits the inefficient utilization of radio frequency spectrum without causing destructive interference to the licensed/primary user communication. In recent past, most of the studies on spectrum sensing are focused on cooperative/multinode detection approaches. However, they are confined to the detection of signals in a single frequency band or narrow band. In order to improve the opportunistic throughput, the CR must sense the signals in multiple bands or wideband. This paper presents a wideband spectrum sensing technique based on energy measurement in each sub band. The simulation result shows that the wideband sensing algorithm detects primary user (PU) signal upto -8dB average SNR with sample size 256, for 8 nodes in cooperation. It is able to detect only upto 0dB average SNR for singe node detection with sample size 64 with detection probabilty P_d ≥ 0.9, and false alarm probability P_f ≤ 0.01. We have also implemented the wideband spectrum sensing in Xilinx Virtex-4 (XC4VSX35-FFG668-10) Field programmable Gate Arrays.