Efficient Spectrum Sensing Scheme under Low SNR

The spectrum sensing problem is one of the most challenging issues in cognitive radio networks. In this letter, a novel spectrum sensing scheme named segmental energy detection (SED) is presented. For the proposed scheme, energy detection will carry out segmentally and finish when reach the predetermined performance. We further derive the largest required observations, average sample number and implementation complexity. Simulation results validate the theoretical derivation, and show the proposed scheme can correctly work. When comparing with the energy detection, the proposed scheme spends less sensing time especially in the low SNR region, and is easy to be implemented.