A new low-cost CFAR detector for spectrum sensing with cognitive radio systems

For cognitive radio technology, energy detection using the fast Fourier transform (FFT) is a common technique for spectrum sensing so that underutilized frequency bands can be opportunistically used while causing no detrimental interference to the primary user. However, traditional detection and false alarm rate analyses of FFT-based detectors require incorporating the characteristics of the propagation environment and the primary user. More specifically, the detection threshold is derived in terms of the statistics of the environment, namely the noise variance, and the probability of false alarm P
_FA
. This paper presents a new decision statistic for a wideband spectrum sensing FFT-based constant false alarm rate (CFAR) detector that has a detection threshold that is independent of the noise variance and windowing sequence. This detector enables multiple frequency bands to be monitored simultaneously and allows for the flexibility to ignore certain FFT bins such as those adversely affected by front-end filtering. For an additive white Gaussian noise (AWGN) channel, the exact probability distribution of the decision statistic is derived as well as the threshold. Implementation of this detector requires neither a calibration period nor real-time numerical threshold estimation algorithms to guarantee a constant false alarm rate. Furthermore, theoretical predictions of P
_FA
were found to be in agreement with laboratory measurements performed with a digital hardware realization of the proposed detector using an field programmable gate array.