Research On The Spectrum Sensing Algorithms In MAC-Layer Of Cognitive Radio System

ABSTRACT:With the rapid development of wireless communication technologies and the sharp growing demand for wireless communication services, spectrum scarcity is becoming more and more prominent. Meanwhile, the traditional fixed spectrum allocation leads to a serious waste of the spectrum resource. Cognitive radio (CR) can dynamically use spectrum, and it is an effective way to solve the problem of spectrum scarcity and improve spectrum utilization efficiency.Spectrum sensing is one of the key technologies of CR. MAC-layer spectrum sensing algorithm is mainly used to control the implementation of spectrum sensing algorithm in physical-layer (PHY-layer). Since the spectrum sensing in PHY-layer has the constraints of the devices and cost, MAC-layer algorithm is required to improve the sensing reliability and efficiency. This dissertation is motivated by the National High Technology Research and Development Program (863 National Major Projects), and mainly study on CR spectrum sensing technology in MAC-layer.Based on CR theory and its spectrum sensing principle, the CR spectrum sensing algorithm in MAC-layer is deeply studied. Firstly, channels are modeled as the alternating renewal process, of which the ON/OFF states are alternately distributed. With the consideration of the missed probability and the interrupted probability, the Optimized Sensing Period algorithm (OSP) is derived in detail with the objective of maximizing the discovery of available opportunities. In order to track the channel changes with time, the Time-varying Optimized Sensing Period algorithm (TOSP) based on the observation window is proposed. MATLAB simulations show compared with the reference algorithm (with fixed initial sensing period), OSP can maximum the discovery the available spectrum opportunities to improve sensing efficiency, but its complexity is greater than the reference algorithm. TOSP can track some channel changes with time, but its complexity is much greater than the reference algorithm. It needs further study about how to choose the length of the observation window.The principles of the Random Sequencing Sensing algorithm (RSS), Channel Utilization Sequencing Sensing algorithm (CUSS) and OFF Probability at the next moment Sequencing Sensing algorithm (OPSS) are introduced in detail in this dissertation. Simulations show OPSS can reduce the Average Switching Latency (ASL) effectively but switching times (M) is a lot. With the consideration of the ASL and M, OFF Probability and Remaining Time at the next moment Sequencing Sensing algorithm(OPRTSS) is proposed. The impact of Tr (retry waiting time) on the algorithms'performance is also discussed. Lots of Simulation results show that compared with OPSS, OPRTSS can reduce ASL but its M grows slightly. To select the appropriate Tr helps reduce the overhead in CR system.