New half-voting cooperative sensing algorithms in cognitive radio

In cognitive radio (CR) networks, hard fusion is widely applied for cooperative energy spectrum sensing, since it requires only one bit to transmit the decision results between sensing nodes and the sensing station. And half-voting is an effective algorithm in hard fusion. In this paper, two half-voting algorithms are proposed to enhance the sensing performance. In the first half-voting algorithm, we adopt linear data fusion with weights based on the SNR of each sensing node. In another algorithm, when the sensing station has no knowledge of each sensing node's SNR, the history decisions are utilized to estimate the weight factors. Analyses and numerical results show that the proposed new half-voting algorithms can significantly improve the sensing performance.