An algorithm for spectrum sensing in cognitive radio using tree-structured filter bank

One of the key requirements in cognitive radio is accurate detection of vacant spectrum bands. In this paper, an energy detector based on a tree-structured quadrature mirror filter bank (TQMFB) is proposed for detecting the center frequencies of the channels (frequency bands). Simulation results are presented to show the trade off between the window size, k, and the error in estimating the center frequency of a specific channel. It is shown that, as the value of FFT size increases, the error in estimation decreases for smaller window sizes and does not change significantly for larger windows. The proposed method is compared with discrete wavelet packet transform (DWPT) based energy detector for complexity and performance. Design examples and simulation show that the proposed method estimates the center frequency of the channels with high accuracies for a specific spectral scenario and signal to noise ratio (SNR).