Eigenvalue-based techniques for continuous sensing model in MIMO CR networks

Many studies on cognitive radio (CR) networks have been carried out to maximize the CR throughput under the restriction of avoiding interference on primary user (PU) network. Aiming at a higher network throughput than in the classic approach, the concept of continuous sensing is used, where secondary receiving nodes are responsible for the spectrum sensing and decision about the presence of the PU while others secondary users (SU) transmit in the same sensed spectrum band. In this paper we compare, in a continuous sensing scheme, the performance of eigenvalue spectrum sensing techniques in a multiple sensor CR network. The soft Akaike information criterion (SAIC) achieves the best performance for detection of the PU signal when the SU signal is strong and the modified maximum eigenvalue detection performs better when the SU signal is week.