Spatial Interpolation of Cyclostationary Test Statistics in Cognitive Radio Networks: Methods and Field Measurements

Abstract

S p ati al I nt er p ol ati o n of C y cl ost ati o n ar y T est St atisti cs i n C o g niti v e R a di o N et w or ks: M et h o ds a n d Fi el d M e as ur e m e nts S a c hi n C h a u d h ari , M e m b er, I E E E, M ar k o K os u n e n , M e m b er, I E E E, S e m u Mä ki n e n, C h a n dr as e k ar a n R a m a n at h a n, J a n O ks a n e n , M e m b er, I E E E, M ar k us L a att a , St u d e nt M e m b e r, I E E E, J ussi R y y nä n e n , M e m b er, I E E E, Vis a K oi v u n e n , F ell o w, I E E E, a n d Mi k k o V al k a m a , S e ni or M e m b er, I E E E

Predictive spectrum occupancy probability‐based spatio‐temporal dynamic channel allocation map for future cognitive wireless networks

“…For a set of measurements, an empirical semivariogram is estimated, which is further fitted to one of the theoretical variogram models . The most commonly used variogram estimator, based on method of moments, was proposed by Matheron; however, it lacks robustness particularly in case of non‐uniform measurement observations.…”

“…The work is extended by Ying et al for TV coverage predictions. Similarly, in the work of Chaudhari et al, a spatial interpolation of cyclostationary test statistic has been carried out in TV band. Although, the cyclostationary test statistic‐based spectrum occupancy estimation is claimed to be a robust and accurate technique but it is non‐blind.…”

“…After digital switchover, the PU activity in TV band is almost static, hence a dynamic band‐like GSM is more favorable for any investigation in a CR scenario. Moreover, the authors in the aforementioned works analyzed the temporal characteristics of the PU activity. In the work of Chaudhari et al, it has been proven that a measurement‐based spatial interpolation approach is more accurate than the DBA approach for constructing an REM.…”

“…For a set of measurements, an empirical semivariogram is estimated, which is further fitted to one of the theoretical variogram models. 12 The most commonly used variogram estimator, based on method of moments, was proposed by Matheron 32 ; however, it lacks robustness particularly in case of non-uniform measurement observations. In this regard, a binned variogram is required and, therefore, we have used a Cressie-Hawkins-based modulus estimator, 33 which is highly robust even against local outliers.…”

Spectrum Sensing of DVB-T2 Signals using a Low Computational Noise Power Estimation

“…In [12][13][14][15], intended at sensing signals such as DVB-T2 signals, correlation-based detectors were proposed by exploiting correlation cyclic prefix or pilot. The exact noise power in these detectors are replaced with the power of received signal which is impracticable because the received signal is highly likely to contain the PUs signals.…”