Discrete Wavelet Packet Based Spectrum Sensing in Cognitive Radio using an Improved Adaptive Threshold

Abstract

Thresholding is a very important technique in the detection of spectrum holes in cognitive radios. It is the benchmark by which a cognitive radio decides if a spectrum hole is present or not in a given frequency spectrum. However, the method by which a cognitive radio computes the threshold for any given scenario can be quite a challenge in terms of accuracy, speed, and efficiency. Fixed thresholds are less cumbersome to implement, but they are characterized by undesirable sensitivity and performance issues in the radio environment in which the cognitive radio operates due to uncertainty in noise level. On the other hand, adaptive thresholds give superior performance to fixed thresholds due to their ability to adjust their computational parameters in response to changes in the cognitive radio environment. In this paper, we present a discrete wavelet packet transform (DWPT) based spectrum sensing in cognitive radio with an improved adaptive threshold in terms of speed. The approach developed in this paper halves the number of computations needed to determine the adaptive threshold for each sub-band channel in a discrete wavelet packet transform. The results obtained show the method is efficient, fast, and dynamic in response to the variations in the cognitive radio environment