Performance analysis of superimposed training-based cooperative spectrum sensing

Abstract

Superimposed training (ST) technique can be used at primary users' transmitters to improve parameter estimation tasks (e.g. channel estimation) at primary users' receivers. Since ST adds the training sequence to the data sequence the total available bandwidth is used for data transmission. The exploitation of the ST sequence in the context of cognitive radio networks leads to a significant increase in the detection performance of secondary users operating in the very low signal-to-noise ratio region. Hence, a considerably smaller number of samples are required for sensing. In this paper, the performance of ST-based spectrum sensing in a cooperative centralized cognitive radio network with soft-decision fusion is studied. Furthermore, a throughput analysis is carried out to quantify the benefits of using ST in the context of cognitive radio for both primary and secondary users.