Cooperative Spectrum Sensing for Centralized Dual Nonparametric Cumulative

For the problem of reducing the detection delay of spectrum sensing in cognitive radio (CR) networks, the work proposed an improved approach of cooperative spectrum sensing (CSS) based on dual nonparametric cumulative sum (DualNCUSUM). In CR network with one primary user, local CR users adopted the NCUSUM to handle the energy observation data, which decreased the local detection delay and the need for prior information of primary user. In order to reduce the bandwidth consumption, 1-bit results of pre-decision are made and transmitted to the fusion center which fused 1-bit results disturbed by ambient noise in reporting channel. The fusion center also used the NCUSUM to accumulate the decision statistics to make the final decision whether the signal of primary user was present. The simulation shows that the proposed DualNCUSUM algorithm has shorter detection delay than the conventional nonparametric CSS approaches while maintaining the false alarm probability at 10%.