Collaborative Spectrum Sensing with Imperfect Gaussian Channel Estimation

The performance of collaborative spectrum sensing for cognitive radio in a lognormal shadowing and Rayleigh fading channel is investigated. Unlike previous works that assume perfect knowledge of the average channel signal-to-noise ratio, this paper considers the realistic case where estimation of the average channel signal-to-noise ratio has error. Furthermore, while previous works have designed and examined estimators based on assuming that the channel samples used in the average signal-to-noise ratio estimators are noiseless, the present work assumes that the channel samples are noisy, as is the case in practical systems. Numerical results show that the probability of missed opportunity decreases as the estimation error decreases, as expected. Perhaps unexpectedly, the results also show that in the presence of noise, there exists a threshold phenomenon for the noise level. Below a particular threshold, the probability of missed opportunity increases as the noise level increases. Yet above this threshold, the probability of missed opportunity decreases as the noise level increases.