Energy detection technique for spectrum sensing in cognitive radio

With unprecedented growth of the subscribers in modern cellular and wireless data communications, there is an acute scarcity of additional bandwidth to meet the ever growing demand. To ease the constraint of additional bandwidth, utilization of the existing system has been a topic of recent interest. Cognitive Radio is a new technique in which the spectral holes in unutilized spectrum are determined to be used for instantaneous communication by secondary users. The Cognitive Radio determines the occupancy of the frequency spectrum observed over a time interval by spectrum sensing methods. Spectrum sensing forms a key front end block of Cognitive Radio systems. This paper deals with the design and simulation of the spectrum sensing algorithm for Cognitive Radio under low SNR scenario.

The Energy Based Spectrum Sensing (EBSS) technique has been identified for its relatively simple implementation. However in the published literature on EBSS, the procedure for the Threshold Energy computation lacks both clarity and defined steps. An attempt has been made to improve the conventional EBSS technique by combining it with the statistical Principal Component Analysis (PCA) technique.

In conventional PCA the ratio of the signal space power to the noise space power do not usually match the actual SNR. This paper proposes a correction factor to the conventional PCA technique. The correction factor is applied to the ratio of decomposed signal space power and the noise space power to equate it to the actual SNR. The noise power obtained through the modified PCA based technique and the chosen value of probability of false alarm determines the threshold energy for the EBSS algorithm. The proposed method which is a combination of PCA and EBSS has been validated for wide range of SNRs, different values of probability of false alarm and frequencies of interest. The correction factor to the PCA and the clearly defined process for Threshold Energy computation invoking the PCA as well as the Radar principles are the contributions of this paper.