Multiband joint detection with correlated spectral occupancy in wideband cognitive radios

Recently, a wideband spectrum sensing scheme referred to as multiband joint detection has been proposed by Quan et al., in which a set of frequency dependent detection thresholds are optimized to achieve the best trade-off between aggregate measures of opportunistic throughput and interference to primary users in cognitive radio (CR) networks. While this scheme shows significant performance gains over benchmark approaches, it employs a frequency-decoupled detector structure that is not optimal in the presence of correlation between subband occupancies, a common situation in CR applications. In this paper, we investigate how a frequency-coupled optimum linear energy combiner (OLEC) structure, recently proposed for single user scenarios, can be integrated into the above multiband joint detection framework to take further advantage of subband occupancy correlation in wideband spectrum sensing. We first analyze the performance of the single-user OLEC and derive expressions for its probabilities of false alarm and missed detection. Using these expressions, we then formulate joint optimization problems for the detection thresholds used by a bank of subband OLECs, with the aim to maximize the aggregate opportunistic throughput under interference constraints. Through numerical experiments with a Markov model of subband occupancy, we show that the use of the OLEC in wideband spectrum sensing can significantly enhance CR performance in terms of these global metrics, when compared to the decoupled multiband processing structure.