Blind spectrum sensing for cognitive radio based on signal space dimension estimation

Based on information theoretic tools, a new spectrum sensing method is proposed in this paper to detect vacant sub-bands in the radio spectrum. Specifically, based on the subspace analysis of the received signal, we present a new method to detect the signal presence in a blind way. We have shown that the analysis of signal dimension can assist blind spectrum sensing procedure. Indeed, we have shown that the slope change, from positive to negative trend, of the signal space dimension curve is representative of the transition from a vacant band to an occupied band (and vice versa). In fact, the number of significant eigenvalues is determined by the value that minimizes the Akaike's information criterion (AIC) and is directly related to the presence/absance of data in the signal. The validation of this new method is based on experimental measurements captured by Eurecom RF Agile Platform operating from 200 MHz to 7.5 GHz. Simulations show good results in terms of spectrum holes detection.