: Spread spectrum frequency hopping systems are widely used in military communications because of their high ability to deal with narrowband interference. The anti-jamming persistence is due to the use of a pre-specified sequences of frequency bands in consecutive time slots, according to a pseudo random table, known only to the corresponding trans-receive pairs. Tracking of these systems is difficult because disruptive users are unable to identify the applied frequency table. In this thesis, the main objective is to extract the frequency table, using a cognitive radio network. To achieve this goal, secondary users monitor all frequency bands in different time intervals to identify those used by primary users, and extract the underlying frequency table. Thus, the identification process is performed by an FH-knowledge-based approach. In this model, it is assumed that only a single primary user is active throughout the band. Hence, at any time only one FH band is in use. To determine this band, all frequency bands are examined by the cooperating users and the measured energies are sent to the base station. In the base station, the data is processed and a decision is made about the busy FH band. In fact, for the linear combination of the available local measurements, the fusion center attachesa specific weight to each secondary user, according to its received SNR. Thus, users with higher SNRs are considered more reliable and are referred with greater weights. The main challenge in this model is the derivation of the optimal weights. In this work we study three criteria to adopt the weights. In the first two attempts, the weights and thresholds are chosen in order to maximize the total probability of detection, and the accumulative probabilities of correct decisions under busy and idle conditions, respectively. In the third one, an FH-knowledge-based approach is pursued to obtain optimal weights by maximizing the probability of correct detection. The performances of these methods are evaluated through computer simulations which show an improvement in these attempts. Moreover, there is no need to calculate thresholds in the latter approach Keywords:cognitiveradio network ,Spread spectrum frequency hopping,Detection,Spectrum sensing,Convex Optimization
