Weak Signal Detection In Cognitive Radio

In recent years, with the rapid development of wireless communication, there are more and more communication equipment. The demand for electromagnetic spectrum is becoming bigger than ever. As a kind of limited resources, electromagnetic spectrum is limited by the fixed spectrum allocation strategy. Spectrum resources is increasingly tense, which has restricted the development of wireless communication technology.In order to solve the problem of spectrum shortage, researchers around the world carried out extensive and in-depth researches, and put forward the concept of cognitive radio in 1999: cognitive radio is a kind of intelligent learning network, by studying the environment information, sense the available spectrum of the environment around, the environment around reduce the conflict to main users as far as possible. Once main users start using spectrum, cognitive radio users must quickly detect the main user, and drop out of the spectrum. Cognitive radio is considered to be the final plan to solve the problem of spectrum shortage. Since then, cognitive radio quickly become the focus in communication field.In order to realize cognitive radio, we first must solve the frequency spectrum access problem of cognitive users, namely the spectrum sensing. Spectrum sensing is the basis and prerequisite of spectrum management to realize spectrum sharing. Cognitive users can only decide whether to access to frequency band of interest when they successful sensed the utilization of the spectrum, and can not interfere with the main user signal. At present commonly used spectrum sensing methods include matched filter detection, smooth circulation characteristics detection, energy detection, etc.Among them, the matched filter detection needs to know priori knowledge about the primary user. Smooth circulation characteristics detection is with high computational complexity. Energy detection is of simple calculation and no priori knowledge about the primary user. So, Energy detection is used most widely.In real environment, noise in cognitive network might be high and signal-to-noise might be low. In this case, spectrum sensing is difficult. This is the so called weak signal detection problem.Energy detection is the most widely used deteciotn algorithmwith low complexity.In this paper, we combine energy detection and sequential detection, the novel deteciotn algorithm overcomes the the shortcoming of energy detection in signal detection. In this paper, we put forward a novel segmentation sequential energy detection method. Traditional sequential detection does sampling while receiving, judging detecting, which increase the computational time. We segment the sampling points and once cognitive user receives a segment of data, it makes a judgment or already. By this method, the proposed method improves the detection performance of energy detection and reduces overhead.Segmented sequential energy detection has the advantages of energy detection, and reduces the sampling points needed for detection performance. However, under the condition of low SNR, segmented sequential energy detection still needs a big number of sampling points. Through further thinking, we introduce the cooperative detection method. The method upload all cognitive nodes on the network to the data fusion center, and make a final judgment in fusion center. Theory and simulation results prove that the sequential energy detection algorithm based on cooperation can greatly improve the performance of energy detection.In the past, researchers usually analysized sensing node under gaussian channel or fading channel. That is assuming the other sensing nodes either completely silent or ignore their existence. In the previous study, we also hypothesize noise as gaussian noise, ignore the influence of other interference in the network. However, in a multi-users environment, interference from other sensing nodes is always there, and will seriously affect the performance of sensing performance. We study a network with numbers of sensing nodes, which randomly distributed in a certain area, and give two methods of false-alarm probability calculation. Finally, according to Neyman-Person criterion, with the acceptable false-alarm probability, we can calculate detection threshold.and detection performance.