Research On Spectrum Allocation In Cognitive OFDM Systems

In the status quo, the shortage of spectrum resources has greatly restricted theadvancement of wireless communication technology. Cognitive radio (CR) technologycan address this disturbing problem effectively. By sensing the wireless environment,CR technology can utilize spectrum resources dynamically and adjust the systemtransmission parameters adaptively, thereby enhancing the system performance,especially the spectrum efficiency. This dissertation combines OFDM technology andCR technology and researches the spectrum allocation in cognitive OFDM systems.Firstly, this dissertation introduces the origin of this project and the researchpurposes of CR technology as well as the current situation of the domestic andinternational research on CR technology.Secondly, the basic theory of spectrum allocation is studied in cognitivesystems, especially on comparing the centralized cognitive systems anddistributed ones and introducing the principles, types and modles of spectrumallocation.Thirdly, it studies the channel state information (CSI) estimation forcognitive OFDM systems in various wireless channels, including channelestimation and signal-to-noise (SNR) estimation. To channel estimation, thisdissertation proposes a ridge regression estimation by using the noise power.Compared with LS estimation, the proposed ridge regression estimation hasbetter performance. To SNR estimation, this dissertation studies variousalgorithms of estimation in AWGN channel and proposes based on cyclic prefix(CP) and data symbols biased SNR estimation and unbiased SNR estimation infrequency selective channels. The proposed two SNR estimators have highaccuracy and low complexity and are very fast.Finally, it researches the wireless channel models in cognitive OFDMsystems. Besides, according to the estimated CSI and the results of spe ctrumsensing, spectrum resourses are allocated to secondary users by means ofLagrange algorithm, greedy algorithm and the improved greedy algorithm. Theperformance of these three algorithms is compared and analyzed by various simulating.