Robust detection analysis of linear cooperative spectrum sensing for cognitive radio networks

Robust detection is employed in this work to cope with uncertainties on the channel gains and the noise power levels in a cognitive radio system based on linear cooperative spectrum sensing. The minimum number of samples required to achieve given false-alarm and missed-detection probabilities is derived as a function of the system parameter uncertainty levels and the nominal SNRs. A lower bound to the received symbol energy required to achieve reliable system operation is derived. This lower bound extends the concept of SNR wall to the case of a cooperative CR system with multiple secondary users. Then, a symmetric CR system scenario is investigated analytically and by numerical simulations. Simple asymptotic results are obtained in this case to relate the minimum number of samples required and the system parameters.