Researches Of Spectrum Sensing Technology Based On Cognitive Radio

With the rapid development of wireless communication, lacking of spectrum sources has been one of the main bottlenecks to limit this development. Cognitive radio, which is an intelligent technology of sharing spectrum, can solve the problem of lacking spectrum sources effectively. It is aware of spectrum environment in multidimensional space, searches for and uses the spectrum that is allocated to primary users but not being used by the primary user at the current time. It can dynamically detect and reuse the vacancy spectrum, make decisions about its radio operating behavior adaptively based on certain learning and decision-making algorithm to exploit the underutilized spectrum in an opportunistic manner. It explores new ways for solving the hard technological of improving spectrum efficiency with the limited spectrum resources.Cognitive radio manages the spectrum resources dynamically, senses and analyzes spectrum in certain regional, finds out the spectrum hole that can be used for communication without interfering with the operations of the licensed networks, and improves spectrum efficiency effectively. To use the technology of cognitive radio, the first and foremost work is sensing the environment of the wireless channels, scilicet spectrum sensing and searching for spectrum holes. Spectrum sensing is a fundamental and key task for cognitive radio. The complicated wireless environment, the various interference level that the licensed system can bear and the variety of the primary signals make the real-time spectrum sensing practically challenging, so the research of spectrum sensing based on cognitive radio is necessary and important.Firstly this thesis introduces the research status at home and abroad on cognitive radio and spectrum sensing based on cognitive radio. Secondly, this thesis deeply researches the concept model, framework, protocol architecture and the key technologies of cognitive radio, discusses the technology of spectrum sensing classified and analyzes various technology of spectrum sensing from the view of principle in detail. Based on the above research and analysis, this thesis focuses on the study of the spectrum sensing algorithm using local transmitter, including two aspects of single-point and multi-point cooperative spectrum sensing algorithm. Accordance with the problems that existing spectrum sensing algorithm needs the noise variance and the prior knowledge of the primary signal, as well as issues such as computational complexity, this paper proposed a blind spectrum sensing algorithm based on the eigenvalue decomposition. Moreover, considering that the existing cooperative spectrum sensing algorithm does not take the credibility of the single-point spectrum sensing into account and needs the prior knowledge of the sensing node in the calculation of the reliability of each node, a cooperative spectrum sensing algorithm based on D-S evidence theory is proposed and implemented.In the research of single-point spectrum sensing technology, firstly, this paper analyzes the advantages and disadvantages of various spectrum sensing algorithm; then, accordance with the problems that spectrum sensing needs the noise variance and the prior knowledge of the primary signal, as well as issues such as computational complexity, this paper proposes a spectrum sensing algorithm based on eigenvalue decomposition. This algorithm is proposed according to the distinguishing property between the signals and noise components in the oversampling received signals. The distinguishing property is that the correlation between signal components in different oversampled channels is high, while the correlation between noise components in different oversampled branches is very low. This paper proposes the received signal model base on the introduction of the multi-channel model of oversampling, and constructes the least square error optimization problem with the parameters of the coefficient for the backward linear prediction of the received signal matrix. Then the papaer discusses the optimization problem in three cases: only existing primary signals, existing primary signals and noise, only existing noise. This thesis introduces the QR decomposition method for solving the optimization problem, obtaines two signal statistics as an indicator of the presence/absence of the primary signal in the received signal based on the combination of linear prediction and QR decomposition of the received signal matrix. When the primary signal is present, the signal statistics that are computed in this paper would differ much more in value from each other than when no primary signal is present. Based on the caclucated statistics, the implementation steps of the algorithm were provided. At last, simulation experiments are built to compare the proposed algorithm and the commonly used energy detector through many respects, including detection probabilities, alarm probabilities and receiver operating curve. Simulations have shown that the proposed spectrum sensing algorithm performs much better than the commonly used energy detector.For the researches of cooperative spectrum sensing, firstly this paper analyzes the problem which the commonly used cooperative spectrum sensing algorithms exist, then aiming at the problem that the fusion center takes the detection node equally, and does not take the credibility of the single-point spectrum detection into account, this thesis proposes the improved cooperative spectrum sensing based on the credibility, and introduces the D-S evidence theory into the fusion rule. This paper studies the proposed cooperative spectrum sensing algorithm based on evidence theory in detail, develops the method of calculating detection credibility under the circumstance that the channel is AWGN and the local detection node uses energy detector, provides the fusion progress to merge the information from each detection node using the D-S fusion rule, and proposes a decision rule to make the final decision using the fusion information. At last, the simulation experiment verifies that this proposed algorithm outperforms the existing cooperative spectrum sensing algorithm remarkably.In conclusion, the main work of this thesis is researching the spectrum sensing based on cognitive radio. The improved single-point spectrum sensing based on eigenvalue decomposition overcomes the problems of the existing spectrum sensing algorithm. It does not need the noise variance and the prior information of the primary signals, and the computational complexity has been decreased. The proposed cooperative spectrum sensing algorithm based on D-S evidence theory provides a new fusion way to fusion the information from each detection node, and the algorithm not only takes the credibility of the single-point spectrum detection into account, but also does not need the prior knowledge of the sensing node. Both the detection performance of the two proposed algorithm have been improved significantly. Due to limited time, there are still many imperfections in the system at present, the author will improve it during afterward works.