Practical issues in spectrum sensing for multi-carrier system employing pilot tones

Abstract

Multi-carrier techniques such as OFDM are widely adopted in the physical layer of nowadays radio communication and broadcasting systems. Therefore in cognitive radio, a robust spectrum sensing techniques for such signals is highly desired. Comparing with traditional Cyclic Prefix (CP) based method, newly reported Time-Domain Symbol Cross-correlation based spectrum sensing (TDSC) method [1][2] shows both advantages in sensing accuracy and robustness against shortness of and channel time spread pollution to CP. This sensing method is able to sense all multi-carrier signals with repeated pilot patterns. In this paper, further validations of TDSC-NP (Neyman-Pearson) and TDSC-MRC (Maximal Ratio Combining) methods are presented with practical issues taken into account, e.g. noise uncertainty, oscillator imperfection and computational complexity. The feasibility of applying TDSC for sensing Filter Bank Multi-Carrier (FBMC) signal is validated via simulation. Both simulated signal and captured DVB-T signal with real-world multi-path channel fading are analyzed and the sensing performance of captured signal agrees with that of simulated signal very well.