Research On Spectrum Sensing Algorithms In Cognitive Radio Based On Throughput Maximization

The traditional spectrum management mechanisms have disadvantages, such as unreasonable resource allocation and low spectrum utilization. The rapidly growing demand of higher speed wireless communication service can not be satisfied, thus the problem of spectrum scarcity become increasingly serious. Cognitive radio, which has been a research hotspot in recent years, can improve the spectrum utilization and alleviate the shortage situation of spectrum resources through accessing idle spectrum in an opportunistic manner. Spectrum sensing is one of the key technologies of cognitive radio which requires fast and accurate identification of idle spectrum. Moreover, it provides spectrum access opportunities for sencondary users under the constraint that primary users are protected from interference. Therefore, the research on spectum sensing algorithms is of great significance.This paper presents the research status and key issues of cognitive radio and spectrum sensing in detail. In addioton, basic spetrum sensing algorithms are analysed systematically and a further research is made on throughput maximizaion based spetrum sensing algorithm. On the basis, a double thresholds spetrum sensing algorithm and a multi-channel spetrum sensing algorithm are proposed.Traditional energy detection using single threshold judgment can not meet both missing detection probability and false alarm probability requirements and demands longer sensing time. To solve the above problems, a double thresholds spectrum sensing algorithm based on throughput maximization is proposed. The algorithm introduces double thresholds to control missing detection probability and false alarm probability respectively. Through the analysis and simplification of the objective function and contraints, throughput maximization problem converts to the joint optimization problem of sensing time and difference value of thresholds. The joint optimization problem is decomposed into two single-variable sub-optimization problems and then is solved by the alternating optimization iteration method. Compared with the traditional method using single threshold, the proposed algorithm reduces the false alarm probability with the requirement of detection probability satisfied and shortens the required sensing time. Hence, the throughput of cognitve user system can be effectively improved.Considering that senseing the whole frequency bands by single cognitve user is limited by the hardware complexity and the cost in practical application, this paper studies the problem of mutli-channel spectrum sensing accomplished by a number of cognitive users from the practical point of view. A multi-channel spectrum sensing algorithm based on throughput maximization is proposed, which optimises sensing time and cognitive user allocation jointly. The algorithm searches the optimal sensing time for each channel and reasonably allocates cognitive user resources among multi-channel. The cognitive user allocation problem is modeled by multiple choice knapsack problem and solved by dynamic programming. Compared with traditional method using fixed same sensing time and average allocation of cognitive users, the proposed algorithm realizes the differential treatment of different channels and leads to an improvement of the throughput of cognitve user system.