AG Kommunikationstheorie


Investigation of model uncertainties in eigenvalue-based spectrum sensing


Reliable signal detection algorithms are an essential element of opportunistic spectrum access. Their role is to facilitate undisturbed communication between license-holding operators. In order to cope with the challenge of deciding, whether a channel is idle or occupied in very low signal-to-noise-ratio (SNR) scenarios, the information from several sensors is taken into account in cooperative spectrum sensing. Due to model uncertainties, which cannot be avoided in practical systems, most detectors, e.g., the energy detector and the cyclostationarity detector suffer from SNR walls, below which robust detection is impossible. In this work, a theoretical analysis of the effects of several model uncertainties on the well-known Maximum-Minimum Eigenvalue (MME) detector is presented. The existence of an SNR wall for the MME detector is shown by evaluating the eigenvalues of statistical covariance matrices. If a transmitted signal arrives at two receivers at different times, a performance reduction and in a special case the non-robustness of the MME detector can be observed. Another uncertainty under consideration is the presence of coloured noise. A lower bound on the SNR wall is derived for the MME detector assuming the noise correlation to be in an uncertainty interval. Numerical evaluations support our claims and prove the presence of an SNR wall.

zurück zur Terminübersicht