Abstract
One promising approach to achieving high precision spectrum sensing in a cognitive radio (CR) system is the energy detector based on Wavelet Packet Transform (WPT). However, energy detector spectrum sensing is affected by the Signal to Noise Ratio (SNR). When the SNR is lower than 5 dB, an energy detector may falsely determine some unoccupied sub-channels as occupied [12]. In this paper, a new noise immune algorithm for spectrum sensing is introduced. This algorithm combines two powerful tools: the wavelet packet analysis and Higher-Order-Statistics (HOS). The use of the proposed technique makes spectrum sensing possible in very low SNR conditions. This allows better utilization of the unoccupied spectrum and high spectrum efficiency usage.