Researches On Spectrum Sensing In Cognitive Radio Networks

With the rapid development of the wireless communication technologies around the world, the contradiction between the limited frequency resource and the requirements of wireless service become even more protruding. Cognitive Radio (CR) is a promising radio processing that can improve the frequency utilization efficiency so to overcome the short of frequency resource problem. It permit the unauthorized wireless devices to temporarily exploit the pre-distributed frequency through certain learning and decision algorithm, given it does not cause interference to the authorized devices, so to improve the spectrum efficiency in frequency, time and space dimensions at the same time. Among the technologies in CR networks, spectrum sensing is an important method that improve the spectrum efficiency through the detection of secondary users with constrained interference to the primary users.We first consider the application of quickest detection in spectrum sensing. Specifically, we proposed hard combing based sequential detection and ordered transmission schemes under different signal-to-noise ratio distribution assumptions. In these schemes, the FC performs hypothesis test each time after it receiving one decision from a SU until the final decision can be made reliably. These schemes can significantly reduce the energy consume without affecting the performance.Then we consider autocorrelation-based spectrum sensing of wideband signals in the presence of correlation between the occupancies of frequency sub-bands in multipath fading environment. To take advantage of the prior knowledge of the occupancy, we develop an alternative detector structure based on affine transmission. This algorithm can exploit the correlation between sub-channels while overcome the high complexity of traditional schemes at the same time.In the end, this thesis points out the later works of this topic, as well as the potential research and application areas of spectrum sensing.