Wide Band Full Duplex Spectrum Sensing with Self-Interference Cancellation–An Efficient SDR Implementation

Abstract

Full duplex Cognitive Radio (CR) achieves more spectral efficiency if it operates in listen and talk mode. It also facilitates to find multiple free available spectrum in a wide band and fast frequency handoff in the case the secondary user uses the frequency occupied by the Primary User (PU). Full duplex CR works using two RF chain with employing the sensing by a dedicated RF chain and transmitting the data with other RF chain. The main challenge involved in full duplex sensing is the self-interference problem, particularly for energy detector based spectrum sensing. Energy detector based spectrum sensing is widely used because of less computational complexity and PU signal knowledge is not required which is an essential feature in wideband spectrum sensing. This work proposes an SDR implementation of full duplex CR using energy detector with self-interference cancellation using Wiener filter. The self-interference signal is predicted by Wiener filter and it is removed in the received signal before apply to the energy detector. The proposed scheme is simulated and validated through an experimental study on SDR platform. The results show the improvement in probability of detection of PU signal. Parallel and pipelined architectures are employed to implement the entire system using the Kintext 7 FPGA on NI USRP RIO 2943R SDR, which result in reduction of sensing time.