Two Algorithms Of Spectrum Sensing For Sparse Signal Analysis

In this paper, we present two improved discrete Fourier transform algorithms of spectrum sensing for sparse signal analysis and give a method of information fusion with multiple sets of samples. The properties of time-shifted samples in aliasing effects are obtained first. Based on the properties, we provide a discrete Fourier transform algorithm, CRSS algorithm. It reduces both the samples and the calculation by half compared with traditional methods. Yet the high performance comes from the loss of relatively omissible information. So the relevant limitation is discussed. Moreover, we present another discrete Fourier transform algorithm, SP algorithm, which is based on radix-4 and the energy superposition. It exploits the zero points of energy and shows notable computational advantages in sparse signals analysis. To lift the restriction on sampling frequency in practice, we also present a way of information fusion which combines information of multiple sets of samples and propose the corresponding solutions for different conditions.