Adaptive energy detection with joint wavelet-frequency domain transform

Traditional energy detection method is susceptible to the interference of low SNR (signal-to-noise ratio) environment in the field of spectrum sensing, and neglecting the localization of available spectrum could also affect the discriminative results of spectrum states. In order to improve the noise sensitivity and discrimination accuracy of energy detection, an adaptive energy detection method by combining wavelet-frequency domain transform was proposed. Signal was decomposed by discrete wavelet packet transform to calculate the sub-band energy; the computational complexity of the adaptive threshold was reduced by combining the norm of energy so as to facilitate comparison with the sub-band energy; the available spectrum was located by fast Fourier transform. And the method was simulated to explore the variable relationship between the adaptive threshold and different performance parameters. Simulation results show that the method has good environmental adaptability and system stability, while the detection error is smaller in different SNR environments. In addition, the frequency domain analysis of the sub-band signal to achieve the reordering of the normalized frequency range, which further improves the accuracy of spectrum sensing.