Efficient Spectrum Searching and Monitoring in Cognitive Radio Network

Two objectives of sensing in cognitive radio (CR) are to detect the primary user (PU) accurately and quickly, which are contradicting objectives. Therefore, many papers try to optimize this tradeoff and find the minimum sensing time which protects the PU. The trends are classified in enhancing false alarm probability (P_f) and detection probability (P_d), optimizing inter-sensing time, in-band sensing (monitoring) time optimization, and out-of-band sensing (search) time optimization. The PU model used in most of these work was a simple two states model (busy/idle renewal process). In this work, we developed a model for the PU in its idle state. The model enables the CR node to benefit from its previous measurements. It assumes that there are multi-idle states, each with specific length and known probability of staying in it. We used this model to find the best sensing time, energy detection threshold, and false alarm probability of the channel being sensed in monitoring. Also, we developed an out-of-band optimization formulation. The formulation finds the best number of channels to sense, the threshold of each channel, the sensing time of each channel, and P_f of each channel such that the PU is protected, the sensing time is minimized, and the CR will find an available channel with very high probability.