Correlation Based Spectrum Sensing with Oversampling and Optimal Weights Selection for OFDM based Networks Coexistence in TVWS

Regulatory bodies recognizes the possibility of an unlicensed and dynamic access to the TV-white spaces. As a consequence, it is expected that heterogenous and independently operated networks will be interested in sharing the same bandwidth within the same service area. Along this direction, this paper addresses the coexistence among multiple heterogeneous secondary networks (SNs) operating in the TV-white spaces by proposing a weighted autocorrelation-based spectrum sensing technique. This is designed for orthogonal frequency division multiplexing signals and is based on an efficient exploitation of the correlation properties of the received signal induced either by the cyclic prefix and by the potential oversampling at the receiver. In particular, the proposed approach, enables the identification of concurrent SNs, belonging to a predefined set, suitably selecting the optimal autocorrelation weights. Furthermore, it is shown that the proposed scheme can operate in a blind mode to detect a potential unknown SN. A theoretical analysis is proposed in order to derive the false alarm probability of the proposed detector to prove that this is a Constant False Alarm Detector. Finally, performance comparisons with alternative schemes by resorting to computer simulations are presented to highlight the advantages of the proposed solution.